VOL 29 NO. 2 APRIL - JUNE 2004 A PUBLICATION OF THE PHILIPPINE ACADEMY OF OPHTHALMOLOGY FOUNDED IN 1969

Transcription

1 A PUBLICATION OF THE FOUNDED IN 1969 VOL 29 NO. 2 APRIL - JUNE 2004 Improving the diagnostic accuracy of the nerve-fiber analyzer The GDx nerve-fiber analyzer: How reliable is it? CCT-IOP correlation among Filipinos 0.4 mg/ml vs. 0.2 mg/ml Mitomycin-C in trabeculectomy Treatment of ocular toxoplasmosis in pregnancy The first reported case of Bardet-Biedl syndrome at UP-PGH Landmark glaucoma trials: what they mean in clinical practice PHISSSN PAPI Cert. No. 291

2 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL - JUNE 2004 EDITOR IN CHIEF Patricia M. Khu, MD, MSc ASSOCIATE EDITORS Romulo N. Aguilar, MD, PhD Marissa N. Valbuena, MD, MHPEd ASSISTANT EDITORS Jessica Marie R. Abaño, MD Ruben Lim Bon Siong, MD Santiago A.B. Sibayan, MD, PhD Jocelyn L. Sy, MD Joseph Anthony J. Tumbocon, MD Harvey S. Uy, MD MANAGING EDITOR Carlos G. Naval, MD EDITORIAL BOARD Romeo V. Fajardo, MD FOUNDING EDITOR, PHILIPPINES Romeo B. Espiritu, MD PHILIPPINES Salvador R. Salceda, MD PHILIPPINES Rossina Lydia A. Ramirez, MD, MHSc PHILIPPINES Roger W. Beuermann, PhD SINGAPORE/USA Donald Tan, MD, FRCS SINGAPORE The PHILIPPINE JOURNAL OF OPHTHALMOLOGY (PJO), the official journal of the Philippine Academy of Ophthalmology, aims to provide a venue for exchange of ideas and information among ophthalmologists and other physicians. It publishes peer-reviewed reports of original clinical and laborator y investigations, epidemiological studies done in the Philippines and other countries, major reviews of specific topics, evaluation of diagnostic and surgical techniques, treatment methods, latest updates, and controversial issues in ophthalmology. Published quarterly, Number One of Volume One is dated January March Entered as a third-class mail matter at the Manila Post Office on February 13, Journal International Standard Serial Number: PHISSN Vol. 29 No.1, March Copyright All rights reserved. Philippine Academy of Ophthalmology Unit 815 Medical Plaza Makati Amorsolo Street, corner De la Rosa Street 1229 Makati City, Philippines Telephone Fax: pjo@pao.org.ph Publishing and Editorial Consultants 5/F King s Court II 2129 Chino Roces Avenue 1231 Makati City, Philippines Tel. Nos to 10 Fax edit@medobserver.com NOTICE TO READERS AND CONTRIBUTORS SUBMISSION Manuscripts should be submitted to the editorial offices of the PHILIPPINE JOURNAL OF OPHTHALMOLOGY (PJO)at the Philippine Academy of Ophthalmology (PAO), Unit 815 Medical Plaza Makati, Amorsolo Street, corner Dela Rosa Street, 1229 Makati City, Philippines. Authors must submit, along with the manuscript, a duly accomplished and signed copyright transfer. No manuscript will be reviewed until the signed copyright transfer is received. The author/s shall pay for the cost of the color separation processing for color photographs included in the manuscript. Manuscripts submitted for consideration are reviewed by editors and other experts in the field. Reviewers are consultants to the editor and are instructed not to discuss the paper with the authors. Reviewers should not cite the manuscript or refer to the work it describes before it has been published. Reviewers should refrain from using the information contained in the manuscript for the advancement of their own work or that of their colleague or institution. SUBSCRIPTION The Journal is published quarterly; subscribers will receive four issues per year with supplements on special topics that the editors deem of interest or significance. PAO members in good standing are automatically enrolled subscribers with no additional fees except for the supplements. Nonmembers may subscribe at the following rates, which include delivery fees (for local subscribers) and online access to the PJO website. Individual Local PhP 1, Foreign USD 30.00* Institutional Local PhP 1, Foreign USD 50.00* Particular issues may be purchased at the following rates per copy subject to availability: Individual Local PhP Foreign USD 10.00* Institutional Local PhP Foreign USD 15.00* *Plus prevailing cost of overseas shipping. DISCLAIMER Statements and opinions expressed in the articles and communications herein are those of the author(s) and not necessarily those of the Editor(s) or the Publisher (Philippine Academy of Ophthalmology). The Editors and the PAO assume no responsibility for any injury and/or damage to persons or property as a matter of product liability or negligence or which otherwise arise from use or operation of any methods, products, instructions, or ideas cited or discussed in any article in the Journal. Although all advertising materials are expected to conform to ethical (medical) standards, the appearance of advertising in PJO does not constitute a guarantee or endorsement by the PJO or the PAO of the quality or value of such product or the claims made for it by its manufacturer. REPRINTS/ ADVERTISING Reprint requests and advertising inquiries may be addressed to the Managing Editor, PHILIPPINE JOURNAL OF OPHTHALMOLOGY, Unit 815 Medical Plaza Makati, Amorsolo Street, corner Dela Rosa Street, 1229 Makati City, Philippines. PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

3 PHILIPPINE JOURNAL OF Ophthalmology ANNOUNCEMENTS PAO 2004 annual confab slated November The Philippine Academy of Ophthalmology (PAO) will hold its 2004 annual convention on November 18 to 20 at the Shangri-La Hotel in Mandaluyong. With the theme Look the Future in the Eye, the convention focuses on pediatric ophthalmology and strabismus to draw attention to the growing number of blind and visually impaired children in the country. They are our hope and future, says PAO President Dr. Marcelino D. Banzon, citing the appropriateness of the theme. Yet, they are also the most vulnerable to the many causes of blindness and visual impairment, which are easily preventable with early screening and management, he adds. Prominent international pediatric ophthalmologists and strabismologists have been invited to shed light on the different ophthalmic disease entities affecting children. Among them are Dr. Kenneth Wright, Dr. Mohamad S. Jaffar, Dr. Ken K. Nischal, Dr. John W. Shore, and Dr. Sonal Farzavandi. Meanwhile, Dr. Tekeyuki Akahoshi and Dr. Abhay Vasavada will share more pearls on phacoemulsification. The opening session will have as speaker civil society leader Victoria Garchitorena, president of Ayala Foundation. Dr. Ma. Dominga B. Padilla, who heads this year s organizing committee, said PAO s 2004 meeting will have many firsts. It will feature two new courses on topics that are becoming more and more important to ophthalmologists ophthalmic photography and computerization of the ophthalmic practice. The annual photography contest will also include a category on digital photography. Also, all registered participants will get CDs containing a compilation of selected lectures. Those wishing to participate may register at the PAO office at: Unit 815 Medical Plaza Makati Amorsolo Street, corner De la Rosa St Makati City, Philippines Telephone Fax: pjo@pao.org.ph Abstracts of presentations will be accepted until September 11. The original and two clear photocopies may be sent to the PAO Secretariat. Those who wish to submit their abstracts on-line may visit the PAO web site at Click on the link to the 2004 Annual Meeting. Joint meeting with AAO eyed for 2005 To mark 60 years of organized ophthalmology in the country The Philippine Academy of Ophthalmology will mark the 60th year of organized ophthalmology in the Philippines with a proposed joint meeting with the Ameri-can Academy of Ophthalmology. Dr. Romulo N. Aguilar, former PAO president and head of the committee undertaking preparations for the celebrations, said the proposal has been accepted by the AAO and a memorandum of understanding is due to be signed soon. Dr. Aguilar said this marks the first time that the AAO is holding a joint meeting with a national ophthalmology organization, having done so in the past only with supranational organizations. The meeting is slated November 28 to December 1 at the Shangri-La Hotel in Mandaluyong. The convention is also expected to draw participants from the Asia-Pacific region. It will serve as a fitting highlight of the PAO s celebration of 60 years of organized ophthalmology in the countr y. Organized ophthalmology began in the Philippines in 1945 with the founding of the Philippine Ophthalmological and Otolaryngological Society. Dr. Aguilar is joined in the anniversary celebration committee by Drs. Salvador Salceda, Alejandro de Leon, Marcelino Banzon, Ma. Dominga Padilla, Carlos Naval, Winston Villar, Teresita Castillo, Reynaldo Santos, Heriberto Guballa, Mar y Rose Pe-Yan, and Ronald Yutangco. 62 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

4 PHILIPPINE JOURNAL OF Ophthalmology VOL. 29 NO. 2 APRIL - JUNE 2004 EDITORIAL 64 Gold standard in glaucoma A perfect gold standard is still not available. As a result, an imperfect but considered best available standard may be chosen as the standard of validity. CASE SERIES 88 Treatment of ocular toxoplasmosis in pregnancy Pregnant state may provoke the recurrence of ocular toxoplasmosis. L I Kump, et al. CASE REPORT 94 Manifestations of Bardet-Biedl syndrome K T Lo, et al. Ocular and systemic manifestations were found in the first reported case of Bardet-Biedl syndrome at the University of the Philippines Philippine General Hospital. OTHERS 107 Instructions to authors Special Issue on Glaucoma ORIGINAL ARTICLES The diagnostic properties of a nerve-fiber analyzer in glaucoma 66 P M Khu, et al. The accuracy of the GDx 400 nerve-fiber analyzer in detecting glaucoma can be improved with the use of corneal compensator to correct for the effect of corneal birefringence. Reliability analyses of the GDx nerve-fiber analyzer 73 P M Khu, et al. The GDx 400 nerve-fiber analyzer has good reliability and can be used to monitor changes in RNFL thickness over time. Correlation of central corneal thickness and Goldmann applanation tonometry among Filipinos 79 M L Lat-Luna, et al. The CCT among Filipinos is normally distributed and is comparable to the distribution obtained by metaanalysis of worldwide data. There is a direct correlation between CCT and IOP among Filipino eyes. 0.4 mg/ml vs 0.2 mg/ml mitomycin-c: A comparison of trabeculectomy outcomes 83 M V Aquino, et al. No significant difference is seen in the outcomes of trabeculectomies using 0.2 mg/ml and 0.4 mg/ml MMC. COMMENTARY The implications on clinical practice of randomized controlled clinical trials in glaucoma 99 P M Khu, et al. Four major glaucoma trials have confirmed the beneficial effects of lowering intraocular pressure in open-angle glaucoma. But they also show that management of glaucoma patients must be individualized. PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

5 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL EDITORIAL What is the gold standard in glaucoma diagnosis? - JUNE 2004 The ideal gold standard should be the best available method of differentiating between those with and without the disease. Currently, the diagnosis of glaucoma is based on definitive changes in the optic-nerve head (ONH) and/or repeatable changes in the visual field tested by standard achromatic perimetr y (SAP). Elevated intraocular pressure (IOP) defined as IOP greater than 21mm Hg was part of the definition of glaucoma more than 10 years ago. But several studies have shown that many patients with ocular hypertension demonstrated no signs of glaucomatous damage during follow-up periods of up to 20 years, even if the condition was left untreated. 1-4 Given the complex relationship between IOP and glaucoma damage, researchers and professional organizations emphasized a new glaucoma concept in which the disease was described as an optic neuropathy, with IOP as only one of several risk factors. 5-7 In a single glaucoma examination, critical ONH evaluation is more sensitive and specific for diagnosing glaucoma than IOP measurement or visual field assessment. ONH evaluation has approximately 85% sensitivity and 90% specificity while visual-field examination has approximately 75% sensitivity and 95% specificity. 8 Examination of either one alone, however, is inadequate because the correlation between structural and functional damage in glaucoma is not exactly linear, especially in the early stage of the disease. Several studies have shown that detectable damage to the ONH and retinal-ner ve-fiber layer (RNFL) is generally present before detectable alteration in the visual field. 9, 10 Approximately 40 to 50% loss of ganglion cells was present before the first defect was detected in the visual field. 11 Correspondence to Patricia M. Khu, MD, MS Department of Ophthalmology and Visual Sciences Philippine General Hospital Taft Avenue, Ermita, 1000 Manila, Philippines Tel: Fax: p_khu@hotmail.com In recent years, several instruments have been developed to assist clinicians in detecting the presence or absence of glaucoma. They were tested for their accuracy against a gold standard. In glaucoma, the choice of a gold standard poses several problems. Each of the tools employed in making the diagnosis of glaucoma involves looking at different aspects of the disease. In evaluating the ONH, one looks for the presence of structural damage. In visual-field testing, one looks at the functional damage. Generally, structural damage in the ONH corresponds to functional damage in the visual field with characteristic glaucomatous defects. In the early stage of the disease, however, there may already be structural damage in the ONH without detectable damage in the visual field (preperimetric stage). Hence, combining both features will increase the sensitivity and specificity in the assessment of glaucoma by standard methods. Var ying sensitivities and specificities have been reported for these new glaucoma instruments, the cause of which is probably an imperfect gold standard. The ideal gold standard should be the best available method of differentiating between those with and without the disease. It should effectively discriminate these groups across the full spectrum of the disease. Many times a perfect gold standard is still not available. As a result, an imperfect but considered the best available standard may be chosen as the standard of validity. Many investigators used the glaucoma experts diagnosis as the gold standard, which was derived from integrating the results of the different glaucoma tests (battery of glaucoma tests) as shown in the article on the Diagnostic Properties of a Nerve-Fiber Analyzer (see pages 66 to 72). Others used the definitive diagnosis of glaucoma derived after several years of following up the patient (natural history of the disease). The latter may be more accurate but more time consuming and difficult because of the long latency of the disease. 64 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

6 Among the recently published randomized controlled trials in glaucoma, most investigators used the findings on SAP as definite diagnosis that glaucoma is present. This may be appropriate for the Advanced Glaucoma Intervention Study (AGIS) 12, 13 or the Collaborative Initial Glaucoma Treatment Study (CIGTS) 14 but for studies involving early glaucoma such as The Early Manifest Glaucoma Trial (EMGT) 15 or the Ocular Hypertension Treatment Study (OHTS), 4 a combination of glaucomatous optic-disc findings and/or visual-field defects is used. Since the publication of the results of several landmark studies in glaucoma, much has been discussed about the role of IOP lowering. We have included in this issue a commentary on several of these randomized controlled studies, summarizing the results, outlining their strengths and weaknesses, and more importantly, analyzing their implications on and application to clinical practice. The Editor in Chief References 1. Schulzer M, Drance SM, Douglas GR. A comparison of treated and untreated glaucoma suspects. Ophthalmology 1991; 98: Lundberg L, Wettrell K, Linner E. Ocular hypertension. Acta Ophthalmol 1987; 65: Kass MA, Gordon MO, Hoff MR, et al. Topical timolol administration reduces the incidence of glaucomatous damage in ocular hypertensive individuals. Arch Ophthalmol 1989; 107: Kass MA, Heuer DK, Higginbotham EJ, et al. The Ocular Hypertension Treatment Study. A randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma. Arch Ophthalmol 2002; 120: Gordon MO, Beiser JA, Brandt JD, et al. The Ocular Hypertension Treatment Study. Baseline factors that predict the onset of primary open-angle glaucoma. Arch Ophthalmol 2002; 120: Collaborative normal tension glaucoma study group. The effectiveness of intraocular pressure reduction in the treatment of normal-tension glaucoma. Am J Ophthalmol 1998; 126: Heijl A, Leske MC, Bengtsson B, et al. Reduction of intraocular pressure and glaucoma progression. Results from the Early Manifest Glaucoma Trial. Arch Ophthalmol 2002; 120: Quigley HA, Schwartz B. Open-angle glaucoma. In: Shingleton BJ, Berson FG, Cantor L, et al, ed. Basic and Clinical Science Course. Section 10. San Francisco: American Academy of Ophthalmology, 1994; Sommer A, Katz J, Quigley HA, et al. Clinically detectable nerve fiber layer atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol 1991; Zeyen TG, Caprioli J. Progression of disc and field damage in early glaucoma. Arch Ophthalmol 1993; 111: Quigley HA, Addicks EM, Green WR. Optic nerve damage in human glaucoma. III. Quantitative correlation of nerve-fiber loss and visual-field defect in glaucoma, ischemic optic neuropathy, papilledema, and toxic neuropathy. Arch Ophthalmol 1982; 100: The AGIS Investigators. The Advanced Glaucoma Intervention Study (AGIS): 7. The relationship between control of intraocular pressure and visual-field deterioration. Am J Ophthalmol 2000; 130: The AGIS Investigators. The Advanced Glaucoma Intervention Study (AGIS): 9. Comparison of glaucoma outcomes in black and white patients within the treatment groups. Am J Ophthalmol 2001; 132; Lichter PR, Musch DC, Gillespie BW, et al. Interim clinical outcomes in the collaborative initial glaucoma treatment study comparing initial treatment randomized to medication or surgery. Ophthalmology 2001; 108: Heijl A, Leske MC, Bengtsson B, et al. Reduction of intraocular pressure and glaucoma progression. Results from the Early Manifest Glaucoma Trial. Arch Ophthalmol 2002; 120: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

7 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL ORIGINAL ARTICLE - JUNE , 2, 3 Patricia M. Khu, MD, MS Edgardo U. Dorotheo, MD 1 Lawrence Tinio, MD 3 Cynthia P. Cordero, MS 4 1, 2, 3 Manuel B. Agulto, MD 1 Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Manila, Philippines 2 Institute of Ophthalmology University of the Philippines Manila Manila, Philippines The diagnostic properties of a nerve-fiber analyzer in glaucoma Potential use as a screening or diagnostic tool 3 Eye Referral Center Manila, Philippines 4 Department of Clinical Epidemiology University of the Philippines College of Medicine Manila, Philippines ABSTRACT Objective To determine the diagnostic properties of the GDx 400 (Laser Diagnostic Technologies, San Diego, CA, USA) nerve-fiber analyzer in normal and in glaucoma patients compared with a battery of glaucoma tests used by glaucoma experts as gold standard. Methods Patients with and without glaucoma underwent a complete eye evaluation, automated perimetry, scanning laser polarimetry with the GDx 400, and opticdisc photography. Two glaucoma experts graded each study eye. Two-by-two tables were constructed for 5 GDx parameters (average thickness, superior average, inferior average, ellipse average, and ellipse modulation) and the GDx number. Receiver operating characteristic (ROC) curves were generated. Correspondence to Patricia M. Khu, MD, MS Department of Ophthalmology and Visual Sciences Philippine General Hospital Taft Avenue, Ermita, Manila 1000 Philippines Tel: Fax: p_khu@hotmail.com Funded by the University of the Philippines-National Institutes of Health (Grant No , awarded to Dr. Paricia M. Khu). Presented in part at the XIX Asia-Pacific Congress of Ophthalmology in Bangkok, Thailand, November Results The study included 355 patients (171 normal, 184 glaucoma). The mean values of the 5 GDx parameters were lower for the glaucoma than for the normal group. The sensitivity and specificity of the GDx 400 were 45.4% and 91.9% if the cutoff level of the GDx number was 71. Ellipse modulation (EM) measures have the best ROC curve with area under the curve of Conclusion The GDx 400 nerve-fiber analyzer is primarily used as a screening tool to detect the presence or absence of glaucoma. Its accuracy can be improved with use of continuous corneal compensator. Key words: Glaucoma, Retinal nerve-fiber layer, Nerve-fiber analyzer, Scanning laser polarimetry The authors have no proprietary or financial interest in any product described in this study. PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

8 SCANNING laser polarimetry (SLP), exemplified by the GDx 400 nerve-fiber analyzer, allows noninvasive quantitative assessment of the retinal nerve-fiber layer (RNFL). Several studies 1-5 showed considerable overlaps in measurements among those suspected with glaucoma and those with early glaucoma. Sensitivities vary from 57% to 82% and specificities from 62% to 92%. 6-7 These differences occurred depending on the cutoff for abnormal used. Other studies 1, 7 used logistic regression analysis and discriminant function and identified several GDx parameters that best differentiate normal from glaucomatous eyes. In this study, the diagnostic properties (sensitivity, specificity, predictive value, likelihood ratio) of the GDx 400 in normal and in glaucoma patients were compared with a battery of glaucoma tests employed by glaucoma experts as gold standard. The receiver operating characteristic (ROC) curve was used to illustrate the relationship between sensitivities and specificities of five GDx parameters and the GDx machine. METHODOLOGY Patients were recruited from the Eye Referral Center (ERC) and the Glaucoma Section of the Department of Ophthalmology, Philippine General Hospital (PGH) between July 2000 and July Included in the study were patients between 30 and 79 years of age with visual acuity of at least 20/40 (6/12) with best correction. Those with significant media opacities that could preclude good scanning images, presence of retinopathy, or high refractive error of greater than 6 diopters were excluded. Only one eye per patient was included in the study. The study complied with the Declaration of Helsinki and was approved by the Ethics Committees of the PGH and the Eye Referral Center. All subjects gave informed consent. Each patient underwent the following examinations: complete eye evaluation, automated threshold perimetry 30-2 (Octopus 101, Bern, Switzerland or Humphrey Field Analyzer I 630, San Leandro, CA, USA), scanning laser polarimetry (GDx 400), optic disc photography (Canon 60UVI fundus camera, Tokyo, Japan). Two glaucoma experts, masked from the GDx results, graded all the eyes included in the study. Eyes were classified as either normal or with glaucoma. The following aided the experts in making the classification: 1. An eye was considered normal if the visual field did not show any characteristic glaucomatous defects on standard achromatic perimetry (SAP), had an intraocular pressure (IOP) of less than 30 mmhg, had a cup-to-disc (CD) ratio of less than or equal to 0.5, and no suspicious glaucomatous damage on the optic-nerve head (ONH). 2. An eye was considered to have glaucoma if any one of the following was present: cupping to the disc margin with associated enlargement of the peripapillary atrophy with or without detectable abnormalities in the visual field on SAP; glaucomatous abnormalities in the ONH such as disc hemorrhage; and abnormalities in the visual field characteristic of glaucoma damage. During the grading sessions, information relating to each study eye was made available (visual-field printouts, IOP, ocular history and medications but no GDx results). All the available optic-disc photos were projected onto a screen at the same magnification for adequate comparison. The classification was done openly with discussions between the two graders. A consensus was arrived at for the diagnosis of each eye. If the disc cupping was more than 0.5, centrally located, with an intact neuroretinal rim, and no observable visual-field defect on SAP, it was classified as normal. Data collected were entered into MS Excel (Microsoft Corporation, Redmond, WA, USA) worksheet and subjected to ROC curve analyses using Statistical Package for the Social Sciences (SPSS) version 11.0 (SPSS, Inc., Chicago, IL, USA). Two-by-two tables were constructed for five select GDx parameters, which included average thickness (AVE), superior average (SA), inferior average (IA), ellipse average (EA), and ellipse modulation (EM). Their sensitivities, specificities, predictive values, and likelihood ratios were obtained at varying cutoff levels for the presence or absence of glaucoma using Epi Info version 6 (Centers for Disease Control and Prevention, Atlanta, GA, USA ). The GDx number (#) 8 represented the overall probability of an eye having glaucoma. It was generated by the machine and derived by integrating all the GDx parameters and characteristics by means of neural network and comparing to an inherent database. Values ranged from 0 to 100; the higher the number, the higher the probability. Values 0 to 30 were considered normal; 31 to 70 glaucoma suspects; and 71 to 100 with glaucoma. Several two-by-two tables were generated at different cutoff levels for the presence or absence of glaucoma. A ROC curve was constructed to determine the best cutoff level. RESULTS Demographic characteristics A total of 355 patients (144 males, 211 females) were included in the study. Mean age was 58 years. Mean visual acuity was 20/25 (6/7.5) (0.9). Mean IOP was 14.5 mm Hg (Table 1). Patients in the glaucoma group were older (p = 0.001), had lower visual acuity (p = 0.003), had abnormalities in the visual field with lower mean sensitivity (p < 0.001), higher mean defect (p < 0.001), and larger loss variance or pattern PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

9 Table 1. Demographic and visual characteristics of patients. Age (years) Mean SD Range Male:female ratio Visual Acuity Mean SD Range Refraction (spherical equivalent) Mean SD Range IOP (mm Hg) Mean SD Range Visual Field Mean Sensitivity (decibels) Mean SD Range Visual Field Mean Defect (decibels) Mean SD Range Loss Variance (decibels) Mean SD Range Vertical Cupping Mean SD Range Horizontal Cupping Mean SD Range Without Glaucoma (n = 171) to 78 66: to to to to to to to to 0.5 With Glaucoma (n = 184) to 78 78: to to to to to to to 0.9 standard deviation (p < 0.001). Many of those with glaucoma were on medication or had glaucoma surgery. GDx parameters: mean values The five GDx parameters (Table 2) showed lower mean values in the glaucoma group compared with the normal group (p < 0.001). Validity analysis Two-by-two tables were constructed for each GDx parameters and ROC curves were generated to determine the best cutoff level for the presence or absence of glaucoma. Table 2. Values of the 5 GDx parameters. GDx Parameters AVE Mean SD Range SA Mean SD Range IA Mean SD Range EA Mean SD Range EM Mean SD Range Without Glaucoma (n = 171) to to to 4.36 With Glaucoma (n = 184) to to to to to 3.98 t-test p-value < < < < < The sensitivities of the five GDx parameters ranged from 62% to 75%, and specificities from 46.8% to 65% (Table 3). The corresponding ROC curves did not show a rapid rise to the upper left corner; instead the curves were closer to the center diagonal line running from lower left to upper right. The positive predictive values ranged from 58.6% to 63.7% and the negative predictive values ranged from 57.1% to 63.6% (Table 3). At best, the GDx parameters were able to predict 60% of those with either a positive or a negative test. Diagnostic properties of the GDx The GDx number was the diagnosis generated by the GDx system and this was compared with the diagnosis given by the glaucoma experts. Different two-by-two tables were generated using several cutoff levels (Table 4). The manufacturer s recommended cutoff level for glaucoma was GDx number 71. The sensitivity obtained on this level was 45.4% with a specificity of 91.9%. High false-negative errors were present. Lowering the cutoff level to 50 would improve the sensitivity to 67.4% and lower the false-negative errors but would also decrease the specificity to 77.8%. Looking at the ROC curve of the GDx (Figure 1), the point nearest the upper left corner has a sensitivity of approximately 68.7% and a specificity of 70%. This was obtained at a cutoff level of 44. The GDx number, which integrates all the GDx parameters, had the best ROC curve when compared with each of the five GDx parameters studied with an area under the curve of (Table 5). The positive predictive value of the GDx number was much 68 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

10 Table 3. Sensitivities and specificities of the 5 GDx parameters using best cutoff level for the presence or absence of glaucoma. Calculations from 2 x 2 tables Sensitivity (%) (95% CI) Specificity (%) (95% CI) Predictive value positive (%) (95%CI) Predictive value negative (%) (95%CI) Likelihood ratio positive (95% CI) Likelihood ratio negative (95% CI) AVE (65) 70.1 (62.9, 76.5) 46.8 (39.2, 54.5) 58.6 (51.8, 65.2) 59.3 (50.5, 67.5) 1.50 (1.25, 1.82) 0.56 (0.43, 0.73) GDx Parameters (Best Cutoff Level) SA (77) 72.7 (63.5, 77.7) 53.0 (48.2, 58.8) 62.0 (54.5, 68.9) 59.1 (51.3, 66.4) 1.51 (1.23, 1.88) 0.64 (0.51, 0.80) IA (87) 75.0 (69.4, 84.7) 54.0 (49.2, 58.5) 63.7 (56.6, 70.3) 63.6 (55.5, 71.1) 1.63 (1.34, 2.0) 0.53 (0.41, 0.68) EA (67) 68.5 (60.1, 75.8) 50.0 (43.2, 58.5) 61.1 (53.9, 67.9) 59.9 (51.7, 67.5) 1.46 (1.21, 1.79) 0.62 (0.49, 0.79) EM (2.11) 62.0 (56.9, 69.6) 65.0 (57.9, 72.0) 59.9 (52.5, 66.9) 57.1 (49.3, 64.7) 1.39 (1.13, 1.71) 0.70 (0.56, 0.87) Table 4. Sensitivities and specificities of the GDx 400 using different cutoff levels of the GDx number for the presence or absence of glaucoma. Calculations from 2 x 2 tables Sensitivity (%) (95% CI) Specificity (%) (95% CI) Predictive value positive (%) (95% CI) Predictive value negative (%) (95% CI) Likelihood ratio positive (95% CI) Likelihood ratio negative (95% CI) < (38.0, 52.9) 91.9 (86.5, 95.3) 85.6 (76.6, 91.6) 61.2 (55.0, 67.2) 6.01 (3.54, 10.37) 0.59 (0.51, 0.67) < (43.1, 57.9) 90.1 (84.3, 93.9) 84.5 (76.1, 90.5) 62.9 (56.4, 68.9) 5.08 (3.21, 8.19) 0.55 (0.47, 0.64) Cutoff Level of GDx Number < (51.0, 65.6) 87.8 (81.7, 92.1) 83.6 (75.8, 89.3) 66.5 (59.9, 72.5) 5.02 (3.31, 7.75) 0.47 (0.39, 0.56) < (53.4, 67.9) 83.0 (76.4, 88.2) 79.4 (71.6, 85.6) 66.4 (59.5, 72.6) 3.59 (2.55, 5.13) 0.47 (0.39, 0.57) < (60.0, 74.0) 77.8 (70.7, 83.6) 76.5 (69.1, 82.7) 68.9 (61.8, 75.3) 3.03 (2.27, 4.11) 0.42 (0.33, 0.52) Table 5. Area under the ROC curve of the 5 GDx parameters 1.00 and the GDx GDx AVE SA IA EA EM GDx # Area % CI higher than the individual GDx parameters studied. It also had a higher positive likelihood ratio (Table 4). DISCUSSION Issues relating to GDx measurements This study shows that the mean values of the five GDx parameters studied were lower in the glaucoma group than in the normal group. But the separation of the glaucoma from the normal was not clear-cut. Overlap of values was present. The ROC curves generated for the five GDx parameters and the GDx number showed curves that were Sensitivity Specificity Figure 1. ROC curve of the GDx nerve-fiber analyzer. PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

11 closer to the diagonal line running from lower left to upper right, indicating that the GDx nerve-fiber analyzer did not adequately discriminate between those with and without the disease. Tests that discriminate well will show that as the sensitivity progressively goes down (the cutoff point is lowered) there is little or no loss in specificity until high levels of sensitivity are achieved. The area under the ROC curve would be much higher and closer to 1. Most overlaps in values were between suspected glaucoma and early glaucoma cases when the study participants were further subdivided into normal, suspected glaucoma, early, moderate, and advanced glaucoma. 9 These results were in agreement with other studies 1-5 showing that the GDx had difficulty differentiating suspected glaucoma from early glaucoma. Issues relating to validity Validity, or accuracy, is the degree to which the results of a measurement correspond to the true state of the phenomenon being measured. A simple way of looking at the relationship between a test s results and the true diagnosis is the two-by-two table. The test is considered to be either positive (abnormal) or negative (normal) and the disease is either present or absent. There are four possible interpretations of test results, two of which are correct and two wrong. The test gives the correct answer when it is positive in the presence of disease or negative in the absence of disease. On the other hand, the test will be misleading if it turns out positive, even when the disease is absent (false positive) or negative when the disease is present (false negative). The results of this study showed that at best the five GDx parameters, when used individually, are able to predict correctly 60% of positive or negative test results. But when the machine is considered as a whole, the positive predictive value can be increased to 77% and the negative predictive value to 69% at a cutoff level of 50. Caution must be exercised in interpreting diagnostic test results. The gold standard used in comparing test results must be known. The choice of a gold standard in assessing the sensitivity and specificity of a new instrument for the diagnosis of glaucoma poses several problems. Each of the tools employed in making the diagnosis of glaucoma involves looking at different aspects of the disease. In evaluating the optic-ner ve head, one looks for the presence of structural damage. In visual-field testing, one looks at the functional damage. Structural damage in the ONH, as exemplified by increased cupping with loss of the neuroretinal rim, corresponds to functional damage in the visual field with characteristic glaucomatous defects. However, early visual-field defects may not be seen on SAP until enough ganglion cells (more than 40%) 10 have been damaged, indicating that structural and functional damage in the early stage of the disease is not linear. Studies have shown that structural damage generally precedes functional damage Comparing findings in the ONH or visual field solely is, therefore, not adequate. When both features were combined and used to make a diagnosis, the sensitivity and specificity rose. Most studies in the last few years incorporated both aspects of the disease in making a diagnosis. Thus, the more appropriate choice of a gold standard by which an instrument can be compared with is to use all available information and tools a battery of glaucoma tests that the glaucoma expert uses to make a diagnosis. The glaucoma expert integrates all test results and arrives at the diagnosis of whether glaucoma is present or not. In this study, the consensus grading of two glaucoma experts was used as the gold standard. They graded the optic-disc features and interpreted the visual-field results. The five GDx parameters generated by the GDx 400 are indices of the double-hump profile of the RNFL and are indirect measures of the RNFL thickness using a scale ranging from 0 to 140. How well the five GDx parameters correlate to the true RNFL thickness (measured in microns) is unknown. The considerable overlap in values for the different groups can be due to any or a combination of the following conditions: 1. Incomplete neutralization of the corneal birefringence. The GDx 400 utilized a fixed corneal compensator. It assumes all eyes have a corneal polarization axis of 15 degrees nasally downward and a corneal polarization magnitude of 60 nm. 12 Recent studies showed that approximately 30% of eyes do not assume this configuration leading to variable results due to incomplete neutralization of the effect of corneal birefringence Structures in the eye that exhibit major birefringence are the RNFL, lens, and the cornea. Incomplete neutralization of the effect of the lens and the cornea can result in increased birefringence that may be incorrectly attributed to that of the RNFL. In some of the eyes in this study, the GDx ner ve-fiber analyzer may be measuring the corneal birefringence in addition to the RNFL birefringence. 2. There is a wide spectrum of glaucoma damage. Structural and functional damages in glaucoma do not correspond all the time Studies have shown that structural damage precedes functional damage; that it takes about 40 to 50% ganglion-cell loss before repeated functional damage is detected on SAP. 10 This discrepancy between structure and function can be seen in the early stage of glaucoma wherein there may already be nervefiber-layer thinning, a structural change, but no visualfield changes yet on SAP. Or in the more advanced glaucoma wherein structural changes are far advanced yet large portions of the visual field may still be intact and vice versa. 70 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

12 3. The normative databases to which the GDx parameters were compared were mostly from multicenter studies done in the United States. The values were largely obtained from Caucasian eyes and may not reflect the normal values for Filipino eyes. The study by Poinoosawmy et al. 4 showed age differences in RNFL thickness between normal white patients and Afro-Carribeans but no data for Asians as the number of subjects was too small. Reported sensitivities and specificities of the GDx400 vary depending on the parameters and cutoff points used. Tjon-Fo-Sang and Lemij 6 used the 2.5 and 97.5 percentiles for the superior/inferior (S/I) ratio obtained from the retardation graph as the cutoff between normal and glaucoma. They reported sensitivity and specificity greater than 90%. Trible et al. 7 used the number of GDx parameters that were abnormal (i.e., 3 or more abnormal parameters considered glaucoma). The study also used the GDx number to denote the probability of abnormality in glaucoma, and had sensitivities of 57% for early, 71% for moderate, and 81% for severe glaucoma when specificity was 89%. Our results are in agreement with the latter lower sensitivities. The GDx number assigned to each study eye by the machine was compared with the diagnosis made by the glaucoma experts. Using the cutoff level of 71, the sensitivity of the analyzer was only 45.4%. This means that the GDx system was able to detect glaucoma in less than half of the patients even though it showed a specificity of almost 92%. If the GDx would be used as a diagnostic tool to detect glaucoma, higher sensitivities are required. To decrease false-negative errors and, therefore, increase the sensitivity at the expense of specificity, the cutoff level of the GDx number may be decreased (Table 4). If the cutoff level were lowered by 10 units, the sensitivity would increase by 13% to 58.5% with only a slight decrease in specificity (87.8%). If the cutoff level were lowered by 20 units, the sensitivity increased to 67.4% and the specificity lowered to 77.8%. There will be more false-positive errors at the expense of increasing the sensitivity level to detect early cases. In glaucoma, detecting the disease early with continued monitoring means a decreased chance of blindness over the long term. A choice has to be made whether to catch all glaucoma cases early by lowering the sensitivity (when used as a screening tool), or to maintain good specificity (when used as a diagnostic tool). This choice will depend on the prevalence of glaucoma in the setting where the GDx machine is to be used. Positive results, even for a very specific test, when applied to patients with a low likelihood of having the disease, will be largely false positives. Similarly, negative results even for a very sensitive test, when applied to patients with a high chance of having the disease, are likely to be false negatives. Thus, the more sensitive a test is, the better will be its negative predictive value, and the more confident the clinician can be that a patient with a negative test result does not have the disease being sought. Conversely, the more specific the test is, the better will be its positive predictive value. One also has to consider the effect of labeling a person glaucoma suspect or with glaucoma, both for economic (cost of follow-up visits and the different glaucoma tests) and psychological reasons (the stigma of the disease and fear of blindness). Once the diagnosis has been made, the question of when to treat comes in. Management of the condition is lifelong and may involve long-term use of expensive glaucoma medications. Hence, in a diagnostic tool, a higher specificity over sensitivity is preferable since falsely labeling a patient with glaucoma has long-term consequences. Limitations of the study Gold standard. Standard achromatic perimetry may not be sensitive enough to detect early glaucoma damage. Newer perimetric tests, such as the short wavelength automated perimetry (SWAP) and the frequency doubling perimetry, 17 have been reported to detect early glaucomatous changes five years before they were evident in SAP. They may replace SAP and if so would certainly be included among the gold standards in detecting the presence or absence of glaucoma. Variable corneal compensator. The GDx 400 used a fixed corneal compensator in correcting for the effect of corneal birefringence. For eyes with corneal polarization axis and magnitude different from that of the machine, increased birefringence may occur that may incorrectly be attributed to that of the RNFL. Newer version of the SLP (GDx Access, Laser Diagnostic Technologies, San Diego, CA, USA) that can neutralize the corneal polarization axis for each eye using continuous corneal compensator, 13 will improve the ability of this technology in detecting early glaucoma. Normative values for Filipino eyes. Obtaining normative values specific for Filipino eyes whereby eyes tested can be compared with will certainly be advantageous since there may be racial differences in the thickness of normal RNFL. Currently, the GDx 400 nerve-fiber analyzer is primarily used as a screening tool to detect the presence or absence of glaucoma. Its accuracy can be improved with use of the continuous corneal compensator to correct for the effect of corneal and lens birefringence. The newer version of the machine is likely to have more precise measurements with less overlap of values, leading to its use as a diagnostic tool to detect early glaucoma. PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

13 References 1. Weinreb RN, Zangwill L, Berry CC, et al. Detection of glaucoma with scanning laser polarimetry. Arch Ophthalmol 1998; 116: Weinreb RN, Dreher AW, Coleman A, et al. Histopathologic validation of Fourierellipsometry measurements of retinal nerve- fiber layer thickness. Arch Ophthalmol 1990; 108: Weinreb RN, Shakiba S, Zangwill L. Scanning laser polarimetry to measure the nerve-fiber layer of normal and glaucomatous eyes. Am J Ophthalmol 1995; 119: Poinoosawmy D, Fontana L, Wu JX, et al. Variation of nerve-fiber-layer thickness measurements with age and ethnicity by scanning laser polarimetry. Br J Ophthalmol 1997; 81: Weinreb RN, Shakiba S, Sample PA, et al. Association between quantitative nervefiber-layer measurement and visual-field loss in glaucoma. Am J Ophthalmol 1995; 120: Tjon-Fo-Sang MJ, Lemij HG. The sensitivity and specificity of nerve-fiber-layer measurements in glaucoma as determined with scanning laser polarimetry. Am J Ophthalmol 1997; 123: Trible JR, Schultz RO, Robinson JC, Rothe TL. Accuracy of scanning laser polarimetry in the diagnosis of glaucoma. Arch Ophthalmol 1999; 117: GDx Nerve-Fiber Analyzer Instruction Manual, San Diego, CA, Khu PM. The diagnostic properties of the scanning laser polarimetry in glaucoma. Masters thesis, University of the Philippines Manila; May Quigley HA, Addicks EM, Green WR. Optic-nerve damage in human glaucoma. III. Quantitative correlation of nerve-fiber loss and visual-field defect in glaucoma, ischemic optic neuropathy, papilledema, and toxic neuropathy. Arch Ophthalmol 1982; 100: Sommer A, Katz J, Quigley HA, et al. Clinically detectable nerve-fiber-layer atrophy precedes the onset of glaucomatous field loss. Arch Ophthalmol 1991;109: Weinreb RN, Bowd C, Greenfield DS, Zangwill LM. Measurement of the magnitude and axis of corneal polarization with scanning laser polarimetry. Arch Ophthalmol 2002; 120: Zhou Q, Weinreb RN. Individualized compensation of anterior segment birefringence during scanning laser polarimetry. Invest Ophthalmol Vis Sci 2002; 43: Greenfield DS, Knighton RW, Feuer WJ, et al. Correction for corneal polarization axis improves the discriminating power of scanning laser polarimetry. Am J Ophthalmol 2002; 134: Polo VP, Larrosa MJ, Pinilla I, et al. Predictive value of short-wavelength automated perimetry: a 3-year follow-up study. Ophthalmology 2002; 109: Polo V, Abecia E, Pablo LE, et al. Short-wavelength automated perimetry and retinal nerve-fiber-layer evaluation in suspected cases of glaucoma. Arch Ophthalmol 1998; 116: Bowd C, Zangwill LM, Berry CC, et al. Detecting early glaucoma by assessment of retinal nerve-fiber-layer thickness and visual function. Invest Ophthalmol Vis Sci 2001; 42: Acknowledgement The author thanks Dr. Alejandro de Leon, Dr. Rossina Ramirez, and Dr. Corazon Ngelangel for their invaluable input to this study. 72 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

14 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL ORIGINAL ARTICLE - JUNE , 2, 3 Patricia M. Khu, MD, MS Edgardo U. Dorotheo, MD 1 Lawrence Tinio, MD 2 Cynthia P. Cordero, MS 4 1, 2, 3 Manuel B. Agulto, MD 1 Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Manila, Philippines 2 Eye Referral Center Manila, Philippines 3 Institute of Ophthalmology University of the Philippines Manila 4 Department of Clinical Epidemiology University of the Philippines College of Medicine Manila, Philippines Reliability analyses of the GDx nerve-fiber analyzer ABSTRACT Objectives The scanning laser polarimetry, exemplified by the GDx 400 (Laser Diagnostic Technologies, San Diego, CA, USA) nerve-fiber analyzer, allows noninvasive quantitative assessment of the retinal nerve-fiber layer. This study determined the reliability of the GDx 400 in taking repeat measurements by different operators and at different sessions in a sample of normal and glaucoma patients. Methods Patients with and without glaucoma underwent a complete eye evaluation, automated achromatic perimetry, scanning laser polarimetry, and optic-disc photography. Retinal nerve-fiber layer (RNFL) measurements were obtained for each group of patients by two trained operators who were masked as to the status of the study eye. Four measurements were obtained for each study eye in the same session and in another session. Reliability measures using intraclass correlation coefficient of five preselected GDx parameters were obtained. Correspondence to Patricia M. Khu, MD, MS Department of Ophthalmology and Visual Sciences Philippine General Hospital Taft Avenue, Ermita 1000 Manila, Philippines Tel: Fax: p_khu@hotmail.com Funded by the University of the Philippines-National Institutes of Health (Grant No , awarded to Dr. Patricia M. Khu) and the Glaucoma Research Foundation Philippines. Presented in part at the XIX Asia-Pacific Congress of Ophthalmology in Bangkok, Thailand, November Results The study recruited 355 patients (171 normal, 184 glaucomatous) ages 30 to 78 years. Intraclass correlation coefficients within operator same session ( ), within operator different sessions ( ), between operators same session ( ), and between operators different sessions ( ) were excellent. The reliability measures for the second session ( ) were higher than for the first session ( ) even for measurements taken by the same operator. Conclusion The GDx 400 nerve-fiber analyzer has good reliability measures and can be used to monitor changes in the RNFL thickness over time. Change in measurements exceeding 20% from baseline should be considered as possible progression. Key words: Glaucoma, Retinal nerve-fiber layer, Nerve-fiber analyzer, Scanning laser polarimetry The authors have no proprietary or financial interest in any product described in this study. PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

15 THE SCANNING laser polarimetry (SLP), exemplified by the GDx 400 (Laser Diagnostic Technologies, San Diego, CA, USA) allows noninvasive quantitative assessment of the retinal nerve-fiber layer (RNFL). This instrument uses a diode laser in the near infrared (780 nm) to obtain RNFL thickness measurements at 65,536 retinal points in a 15-by-15-degree grid centered on the opticnerve head (ONH). 1-2 The polarized light emitted passes through the RNFL, undergoes a phase shift that splits into two beams of different velocities. A detector in the instrument measures the degree of phase shift. This measurement, the difference in velocity, is called retardation and is proportional to the thickness of the RNFL through which the incident light has passed. The scanning laser polarimeter uses a mathematical algorithm developed by the manufacturer eliminating the effect of the cornea and the lens as the polarized light passes through them. The measurement obtained is a measure of relative nerve-fiberlayer thickness. The correlation between retardation measurements and RNFL thickness has been shown previously in experiments with histopathological measurements in postmortem human and monkey eyes. 3 In recent years, several studies have validated the principle of SLP; the normal optic nerves have a typical doublehump configuration when the profiles of the RNFL surface height were plotted two dimensionally. 4-5 The thickest areas of the nerve-fiber bundle are located superiorly and inferiorly, whereas the temporal areas tend to be lowest because of slight tilting of the ONH. Glaucomatous atrophy is indicated as a loss of the double-hump pattern with flattening and lowering of the RNFL surface profile. Repeatability studies of the SLP were done on measurements taken at three different peripapillary locations. 6-7 Coefficient of variation ranged from 3.6 to 4.1% in normal eyes and 5.7 to 10.2% in glaucomatous eyes. Precision was calculated up to 5 microns. Small sample size and failure to stratify the glaucoma group according to severity were evident. Whether the measurements were affected by the different operators who took the scanning images was not determined. The ultimate usefulness of objective measurements using sophisticated tools still requires knowing the amount of variability inherent in their use. When the amount of noise present in the acquisition of these images is known, any change that is beyond the noise level would be considered related to the disease process. The usefulness of an instrument is to detect not only the presence or absence of glaucoma, but eventually to measure the progression of the disease process accurately and reliably over the long term. Thus, this study determined the reliability of the GDx 400 nerve-fiber analyzer. Specifically, we determined the within-operator (intra-observer) and between-operators (inter-observer) variability, as well as the within-session (same visit) and between-sessions (different visits) variability of the GDx in a sample of normal and glaucoma patients. METHODOLOGY This is a cross-sectional study among a sample of individuals with and without glaucoma. RNFL measurements were obtained for each group by two operators on the same day (session 1) and on another day (session 2). Reliability measures for five GDx parameters were obtained. Patients seen at the Glaucoma Section of the Department of Ophthalmology of the Philippine General Hospital (PGH) and at the Eye Referral Center (ERC) were recruited based on the following criteria: age 30 to 79 years and visual acuity of at least 20/40 (6/12) or better with best correction. Those with significant media opacities as to preclude good scanning images, presence of retinopathy, or high refractive error of greater than minus 6 diopters were excluded. Only one eye per patient was included in the study. The study was conducted according to the tenets of the Declaration of Helsinki. All subjects gave informed consent. The study was also approved by the Ethics Committees of the PGH, the University of the Philippines College of Medicine, and the Eye Referral Center. Each patient underwent complete eye evaluation, automated threshold perimetry 30-2 (Octopus 101, Bern, Switzerland or Humphrey HFA I 630, San Leandro, CA, USA), scanning laser polarimetry (GDx 400), optic-disc photography (Canon 60UVI fundus camera, Tokyo, Japan). SLP measurements The patient faced the optoelectronic scan head of the GDx 400 with the pupil in undilated state, and fixated on an external target with the eye not being examined. After the operator has properly focused a ring-shaped target onto the iris and centered the target on the patient s pupil, a live fundus image was seen on the liquid-crystal-display monitor. The intensity of the illuminating laser light was adjusted to achieve appropriate fundus illumination. A complete scan consisting of 65,536 individual retinal locations (256 x 256 pixels) with a field of view of 15 o was obtained. The acquisition time was 0.7 second per image. Immediately after acquiring and storing the data in a personal computer, a computer algorithm calculated the amount of retardation at each measured retinal position and expressed it as the RNFL thickness. A retardation map described the change in the state of polarization (retardation) at each location within the field of view. Processing time was approximately 15 seconds. The RNFL images were obtained by two trained and 74 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

16 experienced operators. They were masked as to the status of the study eye. At the first session, two images were obtained by each of the operators, total of four images for the study eye. Within one week, the second set of images was obtained by the same two operators. The RNFL measurements stored in the computer were recalled for analyses. The retardation (in degrees) was measured within a 10-pixel-wide band located concentrically with the disc margin at 1.7 disc diameters. The optic-disc margin was approximated by a circle or ellipse placed around the inner margin of the peripapillary scleral ring by the experienced operator. The retardation map was divided into four retinal regions: a superior and an inferior region of 120 degrees each, a temporal region of 70 degrees, and a nasal region of 50 degrees. Mean absolute retardation was calculated for the overall peripapillary retina (360 o ), superior (180 o ) and inferior retina (180 o ), temporal (70 o ) and nasal (50 o ) retina. Several GDx parameters were calculated by the computer. For the purpose of this study, the reliability of the following five parameters that give the best sensitivity based on previous studies 8 were evaluated: 1. Average thickness (AVE) the average of all of the thickness measurements in the image; 2. Superior average (SA) the average of the 1,500 thickest points in the superior quadrant; 3. Inferior average (IA) the average of the 1,500 thickest points in the inferior quadrant; 4. Ellipse average (EA) the average of the thickness measurements along the ellipse; 5. Ellipse modulation (EM) the difference between the thickest and thinnest areas along the ellipse. Main outcome measures The reliability of each of the five GDx parameters was determined for the two operators who took the RNFL images. Within-operator (intra-observer) (comparison of images taken by the same operator of the same eye) and between-operators (inter-observer) (comparison of images taken by different operators of the same eye) reliability measures were obtained. Comparison of images taken at two different sessions within one week where there is no expected clinical change was determined. Within-session (comparison of images of the same eye taken on the same visit) and between-sessions (comparison of images of the same eye taken at two different visits) reliability measures were obtained. Intraclass correlation coefficients were obtained to assess the agreement of RNFL measurements under each of the following conditions: a. within-operator (intra-observer) within-session b. within-operator (intra-observer) between-sessions c. between-operators (inter-observer) within-session d. between-operators (inter-observer) between sessions Statistical analysis The intraclass correlation coefficient ρ is defined by the following equation: MS B-pt MS W-pt ρ= MS B-pt + (m 0-1) MS W-pt Where MS B-pt = between-patient mean square difference MS W-pt = within-patient mean square difference m 0 = number of ratings, in this case 2 (sessions 1 and 2) The mean square difference is the sum of the squares divided by the degrees of freedom. ρ estimates the percentage of total variance due to the between-patient component. A high value of ρ suggests the noise of the measurement method is low relative to the total variance in the population. 9 Table 1. Demographic and visual characteristics of patients (n = 355). Age (years) Mean SD Male:female ratio Visual Acuity Mean SD Refraction (spherical equivalent) Mean SD IOP (mm Hg) Mean SD Visual-Field Mean Sensitivity (decibels) Mean SD Visual-Field Mean Defect (decibels) Mean SD Loss Variance (decibels) Mean SD Vertical Cupping Mean SD Horizontal Cupping Mean SD Without Glaucoma (n = 171) : With Glaucoma (n = 184) : PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

17 RESULTS Between July 2000 and July 2002, 355 patients (144 males, 211 females) were recruited into the study. The mean age was 58 years. Mean visual acuity was 20/25 (6/7.5)(0.9). Mean IOP was 14.5 mm Hg. The characteristics of the included patients in each group are shown in Table 1. Only 227 (64%) patients came back for repeat RNFL measurements. Reliability analysis The intraclass correlation coefficient ρ, which measures the consistency or agreement of values within cases, was excellent for the five parameters for within-operator within-session reliability measures (Table 2) and for withinoperator between-sessions reliability measures (Table 3). The EM parameter has slightly lower ρ compared with the other parameters. Good intraclass correlation coefficients (ρ) were obtained for all GDx parameters for between-operators within-session reliability measures (Table 4). The EM parameter has slightly lower ρ compared with the other parameters. The inter-observer same session reliability measures were also higher for the second session than the first session. Between-operators between-sessions reliability measures showed good ρ for all parameters with all values above 0.8 (Table 5). This means that repeated measurements taken by different operators on different days for the same patient were reliable. DISCUSSION Since the GDx nerve-fiber analyzer was introduced, it has been promoted as a screening tool for the detection of glaucoma. Several studies, however, showed considerable overlap of values of the GDx measurements such that there was no clear-cut separation between normal and glaucoma eyes. More recent studies 8, showed lower sensitivity and specificity values for the GDx analyzer than originally demonstrated , 16 Recent validation studies indicated that it was more suited for documenting established glaucoma rather than for detecting early glaucoma. It is for this reason that repeatability studies are needed to determine if the GDx 400 can monitor glaucoma reliably over the long term. Since repeat measures are involved, it is essential that we know the amount of variability present that is related to the process of taking the measurements. Variability in repeat measurements can occur when different operators take the images or when images are taken at different sessions even by the same operator. Our results showed excellent ρ for the five GDx parameters when images were taken by the same operator at the same session (Table 2) or at different sessions (Table 3). Excellent ρ values were also obtained by different operators at Table 2. Within-operator (intra-observer), within-session (same session) reliability measures taken during the first session. GDx Parameters AVE SA IA EA EM Table 3. Within-operator (intra-observer), between-sessions (two different sessions) reliability measures. GDx Parameters AVE SA IA EA EM Table 4. Between operators (inter-observer), within-session (same session) reliability measures taken during the first session. GDx Parameters AVE SA IA EA EM ρ * * intraclass correlation coefficient ** confidence interval * intraclass correlation coefficient ** confidence interval Operator 1 (n = 227) Operator 2 (n = 227) 95%CI ** 95%CI ** Session 1 (n = 355) Session 2 (n = 227) ρ * %CI ** ρ * %CI ** Table 5. Between-operators (inter-observer), between-sessions (two different sessions) reliability measures. GDx Parameters AVE SA IA EA EM ρ * * intraclass correlation coefficient ** confidence interval * intraclass correlation coefficient ** confidence interval Operator 1 (n = 355) Operator 2 (n = 355) 95%CI ** 95%CI ** ρ * the same session (Table 4) and at different sessions (Table 5). Almost all values were 0.8 or better and considered excellent. 9 Hence, the GDx 400 showed that it can reliably take RNFL images in the same eye by different operators at different sessions. Learning effect can also improve the reliability measures; measurements taken during the second session, within one week from the first session where there is no likelihood of a clinical change in the eye, have higher ρ than those of the first session (Table 4). During the first session, each patient underwent multiple tests that lasted two and one half hours. The RNFL ρ * ρ * 95% CI ** PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

18 measurements were obtained toward the end of the session and patients already showed signs of fatigue. The second session, moreover, lasted less than one hour and the only test done was a repeat RNFL measurements. Thus, one way to improve the reliability of the GDx measurements is to ensure that patients know what the procedure involves. One of the recommendations of the Association of International Glaucoma Societies (AIGS) is using digital imaging as a clinical tool to enhance and facilitate assessment of the optic nerve and RNFL in the management of glaucoma. 17 Moreover, automated analyses of the results should use appropriate databases in identifying abnormalities consistent with glaucoma. In recent years, different imaging techniques 5, 6, 8, 10, capable of documenting and quantifying the optic-nerve-head features and the RNFL were developed. These imaging technologies may be complementary and may detect different abnormal features in the same patients. 17 Hence, they should not probably be compared against each other; rather, they should be used to enhance the clinical decision-making of the ophthalmologist in monitoring the disease process. Good repeatability of measurements is a prerequisite for following any change in measurement over time. The good intraclass correlation coefficients obtained for the nerve-fiber analyzer for different operators and at different sessions showed that this machine can be used for repeat measures over time to determine any progression of the disease. Variations in measures can be obtained in the same eye using the same instrument and by the same operator. These variations are called fluctuations and can be found in any eye being measured. Some common examples are the short- and long-term fluctuations found in visual-field tests. For the different GDx parameters, there are also fluctuations in measurements over time. These fluctuations usually consist of the variability measured with regard to the different operators and different sessions. For a change to be considered real, the normal fluctuations in measurements must be exceeded. In this study, if fluctuations in measurements should exceed 20% from baseline, the possibility of disease progression should be considered. Future studies on the long-term use of the GDx machine with follow-up GDx measurements will provide answers as to the amount of noise that must be exceeded to consider the change in measurements as progression of the disease. Possible sources of bias in this study include patient selection and performance bias. Patients included in the study were those who could undergo multiple eye tests for several hours and, therefore, tended to be much younger (mean age of 58 years) than the general population of elderly where most established glaucoma is found. In addition, the older population has more difficulty performing reliable visual-field tests and is also expected to do less well at the GDx 400. The reliability of the GDx test results, however, is influenced less by patient s cooperation, as the testing procedure is much shorter and involves only good fixation. In our study, there was no difference in the reliability measures of the normal (mean age of 55 years) and the glaucoma (mean age of 60 years) patients. All the glaucoma patients in this study have 20/40 (6/12) or better vision. Patients with advanced glaucoma and poorer visual acuity may do less well with greater variability than those shown in this study. Caution in the interpretation of follow-up GDx measurements in the elderly must be exercised even in the presence of variability greater than 20% of baseline since these may still be attributed to the patient. Operators who took the measurements in this study were likely to be giving more time and effort in the acquisition of the images more than they normally do for clinic patients. This study did not look into the variability of using other GDx machines of the same model (variability between machines) nor of a different model such as the GDx Access (an improved version) Prudence dictates that adequate baseline should be obtained for all patients. The optimum number of repeat measurements as baseline has not been determined and is beyond the scope of this study. In the long-term follow-up of glaucoma patients, whenever there is a change in the GDx machine, new baseline studies should be obtained. In summary, the GDx 400 nerve-fiber analyzer has good reliability and can be used to monitor changes in the RNFL thickness over time. Change in measurements exceeding 20% from baseline should be considered as possible progression. References 1. Dreher AW, Reiter K. Retinal laser ellipsometry: a new method for measuring the retinal nerve-fiber-layer thickness distribution. Clin Vision Sci 1992; 7: Dreher AW, Reiter K. Scanning laser polarimetry of the retinal nerve-fiber layer. In: Goldstein DH, Chipman RA, eds: Polarization analysis and measurement. Proc SPIE 1992; 1746: Weinreb RN, Dreher AW, Coleman A, et al. Histopathologic validation of Fourierellipsometry measurements of retinal nerve-fiber-layer thickness. Arch Ophthalmol 1990; 108: Caprioli J. The contour of the juxtapapillary nerve-fiber layer in glaucoma. Ophthalmology 1990; 97: Weinreb RN, Shakiba S, Zangwill L. Scanning laser polarimetry to measure the nerve-fiber layer of normal and glaucomatous eyes. Am J Ophthalmol 1995; 119: Chi QM, Tomita G, Inazumi K, et al. Evaluation of the effect of aging on the retinal nerve-fiber-layer thickness using scanning laser polarimetry. J Glaucoma 1995; 4: Hoh ST, Ishikawa H, Greenfield DS, at al. Scanning laser polarimetry: reproducibility of peripapillary nerve-fiber-layer thickness measurements. Invest Ophthalmol Vis Sci 1997; 38(Suppl): Weinreb RN, Zangwill L, Berry CC, et al. Detection of glaucoma with scanning laser polarimetry. Arch Ophthalmol 1998; 116: Snedecor GW, Cochran WG. Statistical methods. 8th ed. Ames: Iowa State University Press, Choplin NT, Lundy DC, Dreher AW. Differentiating patients with glaucoma from glaucoma suspects and normal subjects by nerve-fiber-layer assessment with scanning laser polarimetry. Ophthalmology 1998; 105: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

19 11. Trible JR, Schultz RO, Robinson JC, Rothe TL. Accuracy of scanning laser polarimetry in the diagnosis of glaucoma. Arch Ophthalmol 1999; 117: Nicolela MT, Martinez-Bello C, Morrison CA, et al. Scanning laser polarimetry in a selected group of patients with glaucoma and normal controls. Am J Ophthalmol 2001; 132: Sanchez-Galeana C, Bowd C, Blumenthal EZ, et al. Using optical imaging summary data to detect glaucoma. Ophthalmology 2001; 108: Choplin NT, Lundy DC. The sensitivity and specificity of scanning laser polarimetry in the detection of glaucoma in a clinical setting. Ophthalmology 2001; 108: Tjon-Fo-Sang MJ, Lemij HG. The sensitivity and specificity of nerve-fiber-layer measurements in glaucoma as determined with scanning laser polarimetry. Am J Ophthalmol 1997; 123: Khu PM. The diagnostic properties of the scanning laser polarimetry in glaucoma. Masters thesis, University of the Philippines Manila, May Weinreb RN, Greve EL. Glaucoma diagnosis: structure and function. Netherlands: Kugler Publications, Zangwill LM, Bowd C, Berry CC, et al. Discriminating between normal and glaucomatous eyes using the Heidelberg retina tomography, GDx nerve-fiber analyzer, and optical coherence tomography. Arch Ophthalmol 2001; 119: Broadway DC, Drance SM, Parfitt CM, Mikelberg FS. The ability of scanning laser ophthalmoscopy to identify various glaucomatous optic-disc appearances. Am J Ophthalmol 1998; 125: Bowd C, Weinreb RN, Williams JM, Zangwill LM. The retinal nerve-fiber-layer thickness in ocular hypertensive, normal, and glaucomatous eyes with optical coherence tomography. Arch Ophthalmol 2000; 118: Zangwill LM, Williams JM Berry CC, et al. Comparison of optical coherence tomography and nerve-fiber-layer photography for detection of nerve-fiber-layer damage in glaucoma. Ophthalmology 2000; 107: Zhou Q, Weinreb N. Individualized compensation of anterior segment birefringence during scanning laser polarimetry. Invest Ophthalmol Vis Sci 2002; 43: Greenfield DS, Knighton RW, Feuer WJ, et al. Correction for corneal polarization axis improves the discriminating power of scanning laser polarimetry. Am J Ophthalmol 2002; 134: Acknowledgment The author thanks Mr. Edgar Q. Buagas, Dr. Roland Puaben, and Ms. Evangeline Marion Abesamis for their invaluable services to the project. 78 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

20 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL ORIGINAL ARTICLE - JUNE 2004 Ma. Margarita L. Lat-Luna, MD Paul I. Guerrero, MD John Vincent Policarpio D. Flores, MD, MS Epi Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Manila, Philippines Correlation of central corneal thickness and Goldmann applanation tonometry among Filipinos ABSTRACT Objective To determine the distribution of central corneal thickness (CCT) among Filipinos and to correlate CCT with intraocular pressure (IOP). Methods A prospective cross-sectional study was performed among Filipino patients consulting at the General Ophthalmology Clinic of the Philippine General Hospital. They underwent a comprehensive eye examination. CCT obtained by ultrasonic pachymetry and IOP by Goldmann applanation tonometry were correlated using linear regression analysis. Factors affecting CCT measurements were analyzed by ANOVA. Results Two hundred twenty two eyes of 112 patients were included in the study. CCT ranged from µm to µm with a mean of µm ±33.8 µm. There was a significant linear correlation between CCT and IOP (r = 0.63). The IOP was noted to rise by 4.3 mm Hg/100 µm CCT. Correspondence to Ma. Margarita Lat-Luna, MD Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Taft Avenue, Ermita 1000 Manila, Philippines Tel ext Telefax margeluna@pacific.net.ph Conclusion The CCT among Filipinos is normally distributed and is comparable to the distribution obtained by metaanalysis of worldwide data. The study also found a direct correlation between CCT and IOP among Filipinos. Key words: Applanation tonometry, Central corneal thickness, Glaucoma, Intraocular pressure The authors have no proprietary or financial interest in any product described in this study. PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

21 ACCURATE measurement of intraocular pressure (IOP) is essential not only in the diagnosis and management of glaucoma, but also in the postoperative monitoring of those who have undergone intraocular surgery. Currently, Goldmann applanation tonometry is the gold standard for clinical measurement of IOP. It works on the principle that the pressure inside an ideal sphere is equal to the force necessary to flatten its surface, divided by the area of flattening. 1 In calibrating this tonometer, Goldmann assumed a standard central corneal thickness (CCT) of 0.55 µm. 1 He predicted that variations from this standard corneal thickness, which he assumed to be rare, would be a source of error in measuring IOP. Currently, there is accumulating evidence that Goldmann s theoretical prediction is correct. Several studies involving simultaneous IOP measurements by applanation tonometry and by manometry have demonstrated that thicker corneas yield an overestimation of IOP by applanation. 2-4 Population-based studies have also demonstrated a positive correlation between CCT and IOP measured by Goldmann applanation tonometry 5, 6 or Tono-Pen. 7 A recent metaanalysis of IOP measured by applanation and CCT, which pooled 133 data sets, revealed a statistically significant correlation between the two. 8 Studies involving patients who have undergone Photorefractive Keratectomy (PRK) 9-13 and Laser in Situ Keratomeleusis (LASIK) 14 have demonstrated a corresponding drop in postoperative IOP with the cornea ablated. Numerous studies have shown that patients diagnosed with ocular hypertension have significantly thicker corneas compared with normal subjects, suggesting that IOP in this group of patients is being overestimated Similarly, patients diagnosed with normal-tension glaucoma have thinner corneas than normal subjects, suggesting that IOP in this group of patients is being 16, 18, 19, 20 underestimated. Given these, several authors have recommended that CCT be considered in interpreting IOP or adjustments 2-5, 7, 11, 13-15, 17, 18, 21, 22 be made in IOP based on CCT. To date, there are no published studies of CCT among Filipinos. It may be inappropriate to apply studies on IOP and CCT done in other countries to the Philippine setting because several studies have demonstrated interracial differences in the range of CCT and factors affecting these values. 7, 8 This study determined the following among Filipinos: the distribution of CCT, the correlation between CCT and IOP, and the association between CCT and other factors such as age, gender, diabetes, thyroid disease, hypertension, trauma, eye surgery, and family history of glaucoma. METHODOLOGY This is a prospective cross-sectional study involving Filipino patients seen consecutively at the General Ophthalmology Clinic of the Philippine General Hospital Outpatient Department from October 17 to 19, Patients over 10 years old with normal IOP, optic-nerve head, and cornea were included in the study. Patients with the following findings were excluded: glaucoma, use of glaucoma medication for the past month, intraocular surgery for the past 3 months, intraocular or retrobulbar tumor, contact lens use for the past 1 month, uveitis, retinal detachment, conjunctivitis, corneal ulcer, corneal dystrophy, and corneal refractive surgery. Patient consent was obtained. Patients included in the study were seen by a single examiner (PG). A case-report form was completed indicating demographic and other factors that could affect IOP and CCT. Proparacaine (Alcaine, Alcon, Fort Worth, TX, USA) was given as a topical anesthetic before IOP and CCT measurements were taken. IOP was measured twice using a Goldmann applanation tonometer (Haag- Streit, Bern, Switzerland) with both measurements within 1 mm Hg. The mean of the two measurements was then recorded. CCT was measured 10 times until all 10 measurements had a standard deviation (SD) of less than 5 using an ultrasound pachymeter (Bio & Pachy Meter AL-2000, Nishi-ku, Nagoya, Japan). Both instruments were disinfected with 70% isopropyl alcohol between each use. The calibration of both instruments was checked daily prior to the start of each study session. Statistical analyses were performed using Epi Info version 6.02c (Centers for Disease Control and Prevention, Atlanta, GA, USA). P values less than 0.05 were considered statistically significant. For all statistical tests, the mean of the ten CCT measurements was used. The correlation between average CCT and IOP, and between CCT and age were determined by linear regression. The effect of gender; family history of glaucoma; diseases such as diabetes mellitus, systemic hypertension, and thyroid disease; trauma; and eye surgrery on CCT was determined by analysis of variance (ANOVA). The right and left eyes were analyzed separately since IOP and CCT measurements between eyes are interdependent. RESULTS The study included 222 eyes of 112 patients of whom 36 were male and 76 were female. Two patients had 1 eye where IOP and CCT could not be measured because of corneal irregularity secondary to trauma. The patients ages ranged from 16 to 86 years with a mean of 52.3 ±17.5 years. Most of the patients were in the 51- to 70-year age group (Figure 1). CCT ranged from µm to µm with a mean of ±33.8 µm (Figure 2). The IOP 80 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

22 ranged from 11 mm Hg to 21 mm Hg with a mean of 16.2 ±2.3 mm Hg. A direct linear correlation was found between CCT and IOP (r = 0.63). The IOP was noted to rise by 4.3 mm Hg/ 100 µm CCT (Figure 3). No significant correlation was found between CCT and age, gender, diabetes, thyroid disease, hypertension, trauma, eye surgery, or family history of glaucoma (p > 0.05). Neither was there a significant relationship between IOP and these factors (p > 0.05). DISCUSSION There is no general consensus on what is the average normal CCT. It has been estimated to range from 518 µm 24 to 580 µm. 21 A recent metaanalysis of measured CCT in normal eyes pooled from 300 data sets worldwide reported a mean of 534 ±31 µm. 8 Our study population had a comparable mean CCT of ±33.8 µm. The CCT values were distributed normally (Figure 2). Our study population also showed a direct correlation between CCT and IOP measured by Goldmann applanation tonometry (r = 0.63). An increase of 4.3 mm Hg per 100 µm of CCT increase was noted. This correlation is lower compared with results reported in manometry studies. Ehlers et al. 2 reported a 5 mm Hg rise for every 70 µm CCT (7.14 mm Hg/100 µm CCT) while Whitacre et al. 4 predicted an error of 3.5 mm Hg for every 70 µm CCT (5 mm Hg/100 µm CCT). Compared with population-based studies by Dohadwala et al., 7 Foster, et al., 5 and Wolfs et al., 6 our computed slope for IOP rise for every µm is steeper. Dohadwala et al. 7 reported only a 2 mm Hg increase per 100 µm CCT. Their study, however, used a Tono-Pen to measure IOP. In contrast, Foster et al. 5 used an optical pachymeter and reported a 1.8 mm Hg/100 µm slope for right eyes and 2.4 mm Hg/100 µm for left eyes. Wolfs et al. 6 reported a rise of 1.9 mm Hg/100 µm. This study, however, only included patients aged Number of Patients Female 13 Male Figure 1. Age distribution of patients (n = 112). Intraocular Pressure (mm Hg) Number of Eyes Age (years) Figure 2. Distribution of central corneal thickness among Filipinos (n = 222). 5 y = 0.043x r = Central Corneal Thickness (µm) Figure 3. Scatterplot of central corneal thickness and intraocular pressure (n = 222) Central Corneal Thickness (µm) PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

23 55 or older (Table 1). Variations in the results obtained may also be partly explained by the differences in the number of eyes examined. Dohadwala et al. 7 reported a significant difference in CCT among Asians, Blacks, and Caucasians. It is likely that interracial variation could partly explain the difference in the magnitude of the relationship observed between other studies and our study. Differences in instrumentation and characteristics of the study population are also possible causes. Other studies have reported that factors such as age, gender, 7, 28, 29 hypertension, 7, diabetes, 7, 20, 28, 33, 34 and family history of glaucoma 28,35 have relationships to IOP and/or CCT. Our study, however, did not show any statistically significant relationship. This study showed that CCT among Filipino patients consulting at the General Ophthalmology Clinic of the Philippine General Hospital is normally distributed and is comparable to the distribution obtained by metaanalysis of worldwide data. The study also found a direct correlation between CCT and IOP. Variation in CCT is a source of systematic error caused by the cornea s resistance to flattening. Further population-based and manometry studies with larger sample sizes are recommended to evaluate the relationship between CCT and IOP as measured by Goldmann applanation tonometry. Such studies could help generate a formula for adjusting IOP measured by applanation to compensate for variations in CCT. References 1. Schottenstein EM. Intraocular pressure and tonometry. In: Ritch R, Shields MB, Krupin T, eds: The Glaucomas, 2nd ed. Missouri: Mosby, 1996; v. 1, chap. 20: Ehlers N, Bramsen T, Sperling S. Applanation tonometry and central corneal thickness. Acta Ophthalmol (Copenh) 1975; 53: Johnson M, Kass MA, Grodzki WJ. Increased corneal thickness simulating elevated intraocular pressure. Arch Ophthalmol 1978; 96: Whitacre MM, Stein RA, Hassanein K. The effect of corneal thickness on applanation tonometry. Am J Ophthalmol 1993; 115: Foster PJ, Baasunhu J, Alsbirk PH, et al. Central corneal thickness and intraocular pressure in a Mongolian population. Ophthalmology 1998; 105: Wolfs RC, Klaver CC, Vingerling JR, et al. Distribution of central corneal thickness and its association with intraocular pressure: The Rotterdam Study. Am J Ophthalmol 1997; 123: Dohadwala AA, Munger R, Damji KF. Positive correlation between Tono-Pen intraocular pressure and central corneal thickness. Ophthalmology 1998; 105: Doughty MJ, Zaman ML. Human corneal thickness and its impact on intraocular pressure measures: a review and metaanalysis approach. Surv Ophthalmol 2000; 44: Abbasogin OE, Bowman RW, Cavanagh HD, McCulley JP. Reliability of intraocular pressure measurements after myopic excimer photorefractive keratectomy. Ophthalmology 1998; 105: Cennamo G, Rosa N, La Rana A, et al. Noncontact tonometry in patients that underwent photorefractive keratectomy. Ophthalmologica 1997; 211: Mardelh PG, Piebenga LW, Whitacre MM, Siegmund KD. The effects of excimer laser photorefractive keratectomy on intraocular pressure using the Goldmann applanation tonometer. Ophthalmology 1997; 104: Table 1. Correlation of central corneal thickness and intraocular pressure in various studies. Study Manometry studies Ehlers, et al. Whitacre, et al. Population-based studies Lat-Luna, et al. Dohadwala, et al. Foster, et al. right eye left eye Wolfs, et al. Increase in IOP per 70 µm Increase in CCT (mm Hg) Increase in IOP per 100 µm Increase in CCT (mm Hg) Munger R, Hodge WG, Mintsioulis G. Correction of intraocular pressure for changes in central corneal thickness following photorefractive keratectomy. Can J Ophthalmol 1998; 33: Rosa N, Cennamo G, Breve MA, La Rana A. Goldmann applanation tonometry after myopic photorefractive keratectomy. Acta Ophthalmol Scand 1998; 76: Emara B, Probst LE, Tingey DP. Correlation of intraocular pressure and central corneal thickness in normal myopic eyes and after laser in situ keratomileusis. J Cataract Ref Surg 1998; 24: Argus WA. Ocular hypertension and central corneal thickness. Ophthalmology 1995; 102: Copt R, Thomas R, Mermoud A. Corneal thickness in ocular hypertension, primary open-angle glaucoma, and normal-tension glaucoma. Arch Ophthalmol 1999; 117: Herndon LE, Choudhri SA, Cox T, et al. Central corneal thickness in normal, glaucomatous, and ocular hypertensive eyes. Arch Ophthalmol 1997; 115: Shah S, Chatterjee A, Mathai M, et al. Relationship between corneal thickness and measured intraocular pressure in a general ophthalmology clinic. Ophthalmology 1999;106: Ehlers N, Hansen FK. Central corneal thickness in low-tension glaucoma. Acta Ophthalmol (Copenh) 1974; 56: Tomlinson A, Leighton DA. Ocular dimensions in low-tension glaucoma. Br J Ophthalmol 1972; 56: Stodmeister R. Applanation tonometry and correction according to corneal thickness. Acta Ophthalmol Scand 1998; 76: Tanaka GH. Corneal pachymetry: a prerequisite for applanation tonometry? Arch Ophthalmol 1998; 116: Ederer F. Shall we count number of eyes or number of subjects? Arch Ophthalmol 1973; 89: Mishima S. Corneal thickness. Surv Ophthalmol 1968;13: Alsbirk PH. Corneal thickness. I. Age variation, sex difference and oculometric correlation. Acta Ophthalmol (Copenh) 1978; 56: Polse KA, Brand RJ, Mandell R, et al. Age differences in corneal hydration control. Invest Ophthalmol Vis Sci 1989; 30: Yee RW, Matsuda M, Schultz RO, Edelhauser HF. Changes in the normal corneal endothelial cellular pattern as a function of age. Curr Eye Res 1985; 4: Armaly MF. On the distribution of applanation pressures: statistical features and the effect of age, sex, and family history of glaucoma. Arch Ophthalmol 1965; 73: Hirvela H, Tuulonen A, Laatikainen L. Intraocular pressure and prevalence of glaucoma in elderly people in Finland: a population-based study. Int Ophthalmol 1994; 18: Dielemans I, Vingerling JR, Algra D, et al. Primary open-angle glaucoma, intraocular pressure, and systemic blood pressure in the general elderly population. Arch Ophthalmol 1995; 102: Shiose Y. The ageing effect on intraocular pressure in an apparently normal population. Arch Ophthalmol 1984; 102: Tielsch JM, Katz J, Sommer A, et al. Hypertension, perfusion pressure, and primary open-angle glaucoma. A population-based assessment. Arch Ophthalmol 1983; 101: Keoleian GM, Pach JM, Hodge DO, et al. Structural and functional studies of the corneal endothelium in diabetes mellitus. Am J Ophthalmol 1992; 113: Larsson LI, Bourne WM, Pach JM, Brubaker RF. Structure and function of the corneal endothelium in diabetes mellitus type I and type II. Arch Ophthalmol 1996; 114: Seddon JM, Schwartz B, Flowerdew G. Case-control study of ocular hypertension. Arch Ophthalmol 1983; 101: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

24 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL ORIGINAL ARTICLE - JUNE 2004 Mario V. Aquino, MD Ma. Margarita L. Lat-Luna, MD John Vincent Policarpio D. Flores, MD, MSEpi Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Manila, Philippines Comparison of outcomes of trabeculectomies using 0.4 mg/ml versus 0.2 mg/ml concentrations of mitomycin-c ABSTRACT Objectives This study compared the outcomes of trabeculectomies using 0.2 mg/ml and 0.4 mg/ml mitomycin-c (MMC) and determined the factors that can predict the postoperative intraocular pressure (IOP). Methods A prospective, randomized, comparative study was performed involving patients undergoing trabeculectomy who were randomly assigned to either 0.2 mg/ml MMC for 4 minutes or 0.4 mg/ml for 2 minutes. The IOP, bleb characteristics, and occurrence of complications were compared. Age and gender of the patients, preoperative IOP, MMC concentration, bleb characteristics, angle status, and age of the surger y were analyzed to determine if they are predictive factors of the postoperative IOP using univariate and multivariate analyses. Results Seventy-four eyes of 68 patients underwent trabeculectomy: 36 eyes were treated with 0.2 mg/ml MMC for 4 minutes and 38 eyes with 0.4 mg/ml MMC for 2 minutes. There was no statistically significant difference in the mean preoperative IOP and postoperative IOP, as well as in the mean percent change in IOP (p = 0.87) between the 2 groups. Univariate and multivariate analyses showed the preoperative IOP (p = 0.02) and the type of filtering bleb (cystic p < 0.001; diffuse p = 0.045) as predictive factors of postoperative IOP. Kaplan- Meier survival curves showed no significant difference between the 2 groups at an average follow-up of 20 weeks. Correspondence to Mario V. Aquino, MD Department of Ophthalmology and Visual Sciences Philippine General Hospital Taft Avenue, Ermita 1000 Manila, Philippines Tel.: ext Telefax: mvaquino@i-manila.com.ph Conclusion There is no significant difference in the outcomes of trabeculectomies using 0.2 mg/ml and 0.4 mg/ml MMC. Preoperative IOP and bleb characteristics are factors predictive of successful filtration surgery. Key words: Trabeculectomy, Mitomycin-C, Intraocular pressure, Filtering bleb The authors have no proprietary or financial interest in any product described in this study. PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

25 THE SURGICAL management of glaucoma is an indispensable alternative in developing countries where there are limitations in medical and laser therapy. Glaucoma filtration surgery fails due to the scarring of the filtering bleb, where fibroblast proliferation from the episclera and Tenon s capsule play a significant role. 1 The use of antimetabolites has improved the success rate of glaucoma filtration surgery. Mitomycin-C is an antibiotic with antineoplastic activity, inhibiting the proliferation of fibroblasts. At present, there is no known consensus as to the ideal concentration and exposure time of mitomycin-c The concentrations used in different studies ranged from 0.02 mg/ml to 0.5 mg/ml. The most commonly used concentrations are 0.2 mg/ml and 0.4 mg/ml. Exposure time varies from 30 seconds to 5 minutes, with the latter as the most common. This study compared the outcome of trabeculectomies using 0.4 mg/ml MMC (Mitomycin C Kyowa, Tokyo, Japan)for 2 minutes as against 0.2 mg/ml MMC for 4 minutes in terms of IOP lowering, characteristics of the filtering bleb, and incidence of complications. It also determined the factors that can predict the postoperative intraocular pressure (IOP). METHODOLOGY Patients seen consecutively at the Glaucoma Service of the Philippine General Hospital (PGH) between January and June 1997 were evaluated. Those who had uncontrolled IOPs, cup-disc ratio of 0.8 or worse, and on maximum tolerated medical therapy were eligible for inclusion into the study. Patients who had systemic contraindications to surgery, who could not comply with the follow-up schedule, or who refused to give their consent were excluded. Based on the clinical diagnosis, patients requiring a filtering procedure were classified as either low risk for surgical failure (e.g. primary glaucomas) or high risk for surgical failure (e.g. congenital and developmental glaucomas, neovascular glaucoma, previous surgeries, failure of previous trabeculectomy, traumatic glaucoma, previous cataract surger y, secondar y glaucoma, inflammatory glaucoma). 18 Using a table of random values, each patient was assigned to either the 2 minute exposure of 0.4 mg/ml MMC or the 4 minute exposure of 0.2 mg/ml MMC. The patients and the surgeons were masked as to which MMC concentration would be used. A trabeculectomy protocol approved by the consultants of the Glaucoma Service was used. After peribulbar anesthesia was given, a corneal bridle using vicryl 8-0 (Ethicon, Johnson & Johnson International, New Brunswick, NJ, USA), was sutured at 12:00. A superior limbal-based conjunctival flap 10 mm posterior to the limbus was created. After hemostasis and clearing of the episclera, a 3x3 mm triangular scleral flap of half thickness was fashioned with the apex of the flap at 12:00. The flap was dissected anteriorly up to the limbal gray zone. A 3mm cut cotton sponge soaked in MMC was applied under the scleral flap. Only the assisting resident was aware of the actual exposure time and was in charge of removing the cotton sponge from the surgical field. The scleral bed was irrigated with 40 ml of balanced salt solution. A paracentesis was done at the temporal limbus and its patency tested. The anterior chamber was entered using blade 11 knife and a sclerostomy was created with 0.75 mm Kelly punch (Storz, St. Louis, MO, USA). Peripheral iridectomy was done with Vannas scissors. The anterior chamber was reformed and the filtering of fluid checked. Three 10-0 nylon sutures secured the scleral flap snugly to the scleral bed; fluid was seen to egress after lightly dabbing on the incision site with a cotton pledget. The Tenon s capsule and conjunctiva were sutured continuously and closed separately with nylon 10-0 suture. All surgeries were performed by the glaucoma fellow and the 7 senior residents who were rotating in the Glaucoma Service during that period. The IOP before surgery and on the most recent followup were compared. The main outcome measure was the change in IOP. The filtering surgery was considered successful when there was at least 30% reduction of IOP from baseline. Secondary outcome measures were the type of filtering bleb and incidence of complications. The type of filtering bleb on the most recent follow-up was classified: cystic, if the bleb was localized, elevated, with cystic changes of the overlying conjunctiva; diffuse, if the border of the bleb was ill-defined, elevated with no visible changes in the overlying conjunctiva; vascular or congested, if the bleb was localized, elevated, and surrounded by ropy blood vessels; or flat, if the bleb is not elevated, difficult to discern (not visible), and there were no changes to the overlying conjunctiva. The outcome assessors were masked as to the patient s group assignment. Univariate and multivariate analyses using Cox proportional hazard modeling were done to determine the effects of factors such as age and gender of the patient, angle status, preoperative IOP, MMC concentration and duration in minutes of MMC application, bleb characteristics, and age of the surgery or duration of follow-up on the postoperative IOP. For variables age and sex, the unit of analysis was based on the patient for person-based covariates. For the other variables studied, the unit was based on the eye for eye-based covariates. Kaplan-Meier survival curves were determined for the two treatment groups to detect any difference in IOP change over the period of follow-up. 84 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

26 Table 1. Distribution of patients according to risk of surgical failure and MMC concentration. Diagnosis Primary open-angle glaucoma Chronic angle-closure glaucoma Intermittent angle-closure glaucoma Acute angle-closure glaucoma, refractory Neovascular glaucoma Failed trabeculectomy Developmental glaucoma s/p Cataract extraction (ECCE) with PCIOL s/p Combined surgery Axenfeld-Rieger syndrome Secondary angle-closure glaucoma Previous intraocular surgery (other than cataract surgery) TOTAL Table 2. Comparison of mean preoperative and postoperative IOPs in eyes treated with 0.2 mg/ml and 0.4 mg/ml MMC. Treatment Group 0.2 mg/ml 0.4 mg/ml Preoperative a p = 0.002, b p = 0.038, c p = Postoperative a b Mean IOP Reduction (mm Hg) c Mean % change Table 3. Distribution of types of filtering bleb and mean IOP change. Bleb Type Cystic Diffuse Vascular Flat Frequency mg/ml MMC 22 (61%) 4 (11%) 6 (17%) 4 (11%) 0.4 mg/ml MMC 23 (61%) 5 (13%) 5 (13%) 5 (13%) Risk Low Low Low Low High High High High High High High High Mean IOP Change MMC Concentration 0.2 mg/ml 0.4 mg/ml 95%Confidence Interval Lower Upper RESULTS A total of 74 eyes of 68 patients (46 females, 22 males; mean age of 51 ± 2.2 years) underwent filtering surgery between January and June Forty-three (58.1%) of these eyes were right eyes. Forty-six eyes were classified as low risk and 28 eyes as high risk (Table 1). The two MMC concentrations studied were evenly distributed between the two surgical risk groups. Thirty-six eyes were treated with 0.2 mg/ml MMC while 38 eyes were treated with 0.4 mg/ml MMC. The mean preoperative IOP was mm Hg and the mean postoperative IOP was mm Hg (Table 2). The mean IOP change for both groups from preoperative levels was mm Hg. There was no statistically significant difference (p = 0.87) in mean IOP reduction between the two groups postoperatively (Table 2). The average percent change in IOP was for both groups. Fifty-five eyes (74.3%) achieved a 40% reduction Table 4. Factors predictive of postoperative IOP (univariate analysis). Factors Studied Age of patient Preoperative IOP Diffuse bleb Vascular bleb Cystic bleb Mitomycin Neovascular glaucoma Duration of follow-up Table 5. Factors predictive of postoperative IOP (multivariate analysis). Factors Studied Mitomycin C Age of patient Diffuse bleb Vascular bleb Cystic bleb Angle closure Neovascular glaucoma Preoperative IOP *odds ratio Hazard Ratio Hazard Ratio % Confidence Interval Lower Upper % Confidence Interval Lower Upper /Exp (B) * p Value <0.001 < < p Value < from baseline IOP. Fifty-eight eyes (78.4%) achieved a 30% reduction from baseline IOP. The mean follow-up period was weeks. The most common bleb in both treatment groups was cystic (Table 3). There was no difference in the type of bleb between the 2 groups. The different bleb types also showed differences in postoperative IOP. Flat bleb had the lowest IOP change of only mm Hg compared to mm Hg for diffuse bleb, mm Hg for vascular bleb, and mm Hg for cystic bleb (Table 3). Complications noted were few: 3 cases of hypotony (2 in 0.2 mg/ml and 1 in 0.4 mg/ml), 4 cases of hyphema (1 in 0.2 mg/ml and 3 in 0.4 mg/ml), 1 case of flat anterior chamber (0.4 mg/ml) and 2 cases of choroidal effusion (0.2 mg/ml). There was no difference in the incidence of complications between the two groups. Univariate analysis showed that preoperative IOP and the different types of bleb are significant predictors of IOP after filtering surgery (Table 4). In multivariate analysis, only preoperative IOP, cystic and diffuse blebs are significant predictors of postoperative IOP (Table 5). There is no evidence that MMC concentration (either the 0.2 mg/ml or the 0.4 mg/ml concentration), angle configuration, patient age, patient gender, age of the surgery in weeks are predictors of IOP after filtering PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

27 Survival Probability Follow-up Time (weeks) Figure 1. Kaplan-Meier survival curve of 0.2 mg/ml and 0.4 mg/ml MMC. MMC 0.4 mg/ml 0.2 mg/ml surgery (Table 4). Linear regression analysis showed that for every 1 mm Hg IOP rise at baseline, there is a corresponding mm Hg reduction in IOP during the postoperative follow-up. The regression model has an R 2 of 82%, indicating that the variability of the IOP after filtering surgery is explained by the identified independent variables in the model 82% of the time. Figure 1 shows the survival curves of the two treatment groups. The difference in survival between the groups is not statistically significant. DISCUSSION Mitomycin-C is one of the most frequently used antifibrotic agents in glaucoma filtration surgery. In conditions where the incidence of bleb scarring is common, resulting in a much lower success rate of trabeculectomy or a higher reoperation rate, use of an antifibrotic agent on a routine basis is justified. In the Philippines, the most common concentrations of MMC used are 0.2 mg/ml and 0.4 mg/ml, with exposure time varying from 2 to 5 minutes. There are two factors to consider when using antifibrotic agents in glaucoma filtration surgery: the concentration of the agent and the duration of application. The strength of MMC application can be varied by either increasing the concentration, usually 0.4 mg/ml over 0.2 mg/ml, or increasing the duration of application to as long as 5 minutes. The total dose of the application can be calculated as concentration multiply by duration of exposure. In this study, we opted for keeping the total dose similar (0.4 mg/ml x 2 minutes versus 0.2 mg/ml x 4 minutes) in order to predict the factors affecting IOP control postoperatively with minimal effect of confounding factors. We compared a high MMC (0.4 mg/ml) concentration applied for a shorter duration (2 minutes) to a lower MMC (0.2 mg/ml) concentration applied for a longer duration (4 minutes) to prevent serious postoperative complications such as scleral necrosis, which has been reported when higher doses were used. Several investigators studied the effects of Mitomycin-C by either using a constant concentration with varying exposure time, 8, 14 a constant exposure time but varying MMC concentrations, 6, 7, 9 or variable concentrations and exposure time. 4, 11 These studies showed no significant difference in the success of IOP control between the various MMC concentrations (0.2 mg/ml to 0.5 mg/ml), except for those with very low concentrations of 0.02 mg/ml 6 to 0.1 mg/ml 7 where the success rate was significantly lower. Varying the exposure time (2 to 5 minutes) also did not show any significant difference among the different groups. Our study similarly showed no significant difference in IOP lowering between the two groups studied at an average follow-up of 20 weeks (Table 3). The survival curves of the two treatment groups were also similar (Figure 1). However, there is a possibility that had the observation period been prolonged, there may result a significant difference between the two, especially for the period beyond 35 weeks. The incidence of complications in this study was similar between the two treatment groups. Several studies reported a higher incidence of postoperative hypotony 7 and hypotony maculopathy 6, 8 in those treated with higher concentrations of MMC and longer duration of exposure. Other studies 9, 11 reported no significant difference. Our study had only 3 cases of hypotony, 2 of which occurred in the 0.2 mg/ml concentration. The similarity of the total dose in the 2 treatment groups in this study may account for the no difference in the incidence of hypotony. Had the duration of exposure been varied, such as to 4 or 5 minutes for the higher concentration MMC especially in the low-risk group, the occurrence of hypotony may increase dramatically. The univariate analysis (Table 4) of the predictive factors for postoperative IOP showed that preoperative IOP and the type of filtering bleb were the only significant predictors. Patient factors such as age and gender were not predictors indicating that MMC use during glaucoma filtration surgery is independent of the patient. Angle status or the type of glaucoma (high risk versus low risk of surgical failure) was not predictive, indicating that MMC use is independent of glaucoma type. The concentration of MMC and the duration of application were also not predictors in this study. Since the total dose of the two treatment groups was similar, the effect of MMC concentration and duration of application may have been negated. Varying the total dose as in varying the exposure time of the higher concentration MMC may have different effects on the postoperative IOP. In the multivariate analysis (Table 5), preoperative IOP still remained a predictor of postoperative IOP. The level of preoperative IOP had an effect on postoperative IOP 86 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

28 or success of the filtering surgery; the higher the preoperative IOP, the greater likelihood that the filter will not fail. This is because of the definition of success rate used in this study. A higher preoperative IOP would result in a higher chance of attaining a 30% reduction of IOP. In the Cox model, for every 1 mm Hg IOP rise at baseline, there was a corresponding or 1 mm Hg reduction in IOP during the postoperative follow-up. In the type of filtering bleb, only diffuse and cystic blebs remained predictive in the Cox model. The more cystic the bleb (also true for diffuse), the lower the postoperative IOP. This was further supported by the mean IOP change for each type of bleb. Most of the blebs in this study were cystic (61%) with the lowest mean IOP change of 22.5 mm Hg, followed by diffuse bleb (Table 6). The vascular bleb, even though it had a large mean IOP change postoperatively, was no longer predictive of a successful filter. This could be due to the shorter follow-up (ranging from 4 to 12 weeks) occurring mostly in those with a vascular bleb. In the Cox model, an eye with a diffuse bleb has 7 times less chance of failure than a flat bleb. Eyes with vascular or cystic blebs have 1.73 and 24.4 chances of increase in probability of success, respectively, compared with a flat bleb (Table 5). Furthermore, the model showed that for every 10 mm Hg reduction from preoperative IOP, the chance of success is increased by 3.35 times (p = 0.02). The regression model has an R 2 of 82%, indicating that the variability of the IOP after filtering surgery could be explained by the identified independent variables included in the model 82% of the time. Other factors, not taken into consideration in this study that may explain the other 18% of the variability are varying exposure time of the MMC concentration, varying follow-up periods, and different surgeons performing the surgery. These factors may affect the ultimate outcome in the differences between using 0.4 mg/ml over 0.2 mg/ml of MMC. Future studies should address these issues, preferably the same surgeon and the same but longer follow-up period for all patients and varying the exposure time. In summary, there is no significant difference observed in the outcomes of trabeculectomies using a 0.4 mg/ml of MMC at 2 minutes versus 0.2 mg/ml of MMC at 4 minutes over an average follow-up period of 20 weeks. Preoperative IOP and type of filtering bleb are factors predictive of a successful filtration surgery. References 1. Collignon-Brach J. Mitomycin-C in glaucoma surgery. Bull Soc Belge Ophtalmol 1993; 247: Annen DJ, Sturmer J. Follow-up of a pilot study of trabeculectomy with low dosage mitomycin C (0.2 mg/ml for 1 minute): Independent evaluation of a retrospective nonrandomized study. Klin Monatsbl Augenheilkd 1995; 206: Bazin S, Williamson W, Poirier L, et al. Trabeculectomy with mitomycin-c: Study apropos of 30 cases. J Fr Ophthalmol 1995; 18: Cheung JC, Wright MM, Murali S, Pederson JE. Intermediate-term outcome of variable-dose mitomycin-c filtering surgery. Ophthalmology 1997; 104: Costa VP, Comegno PE, Vasconcelos JP, et al. Low-dose mitomycin-c trabeculectomy in patients with advanced glaucoma. J Glaucoma 1996; 5: Kitazawa Y, Suemori-Matsushita H, Yamamoto T, Kawase K. Low-dose and highdose mitomycin trabeculectomy as an initial surgery in primary open-angle glaucoma. Ophthalmology : Lee JJ, Park KH, Youn DH. The effect of low- and high-dose adjunctive mitomycin C in trabeculectomy. Korean J Ophthalmol 1996; 10: Megevand GS, Salmon JF, Scholtz RP, Murray AD. The effect of reducing the exposure time of mitomycin-c in glaucoma filtering surgery. Ophthalmology 1995; 102: Mietz H, Krieglstein GK. Short-term clinical results and complications of trabeculectomies performed with mitomycin-c using different concentrations. Int Ophthalmol 1995; 19: Mirza GE, Karakucuk S, Dogan H, Erkilic K. Filtering surgery with mitomycin-c in uncomplicated (primary open angle) glaucoma. Acta Ophthalmol Copenh 1994; 72: Neelakantan A, Rao BS, Vijaya L, et al. Effect of the concentration and duration of application of mitomycin-c in trabeculectomy. Ophthalmic Surg 1994; 25: Ramakrishnan R, Michon J, Robin AL, Krishnadas R. Safety and efficacy of mitomycin-c trabeculectomy in southern India: A short-term pilot study. Ophthalmology 1993; 100: Rozsival P, Liehneova I. Initial experience with the use of mitomycin-c in antiglaucoma filtering surgery. Cesk Oftalmol 1994; 50: Schnyder CC, Bernasconi O, Mermoud A, Faggioni R. Comparative study of administration time of mitomycin-c in trabeculectomy: 2.5 or 5 minutes? Klin Monatsbl Augenheilkd 1995; 206: Singh J, O Brien C, Chawla HB. Success rate and complications of intraoperative 0.2 mg/ml mitomycin-c in trabeculectomy surgery. Eye 1995; 9: Wu L, Yin J. The effect of mitomycin-c on filtration surgery of glaucoma with poor prognosis. Chung Hua Yen Ko Tsa Chih 1996; 32: Zhang M, Wei H, Mai C, et al. Intraoperative use of mitomycin in trabeculectomy. J Tongji Med Univ 1996; 16: Singh G, Kaur J, Dogra A. Intraoperative mitomycin-c in complicated glaucomas. Indian J Ophthalmol 1993; 41: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

29 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL CASE SERIES - JUNE 2004 Leila I. Kump, MD Ian Paredes, MD C. Stephen Foster, MD, FACS Massachusetts Eye and Ear Infirmary Harvard School of Medicine Boston, Massachusetts, USA Treatment of ocular toxoplasmosis in pregnancy ABSTRACT Objectives To describe the course of ocular toxoplasmosis during pregnancy. Methods This is a retrospective, noncomparative case series of four pregnant women who were treated for ocular toxoplasmosis during pregnancy. Results All of the participants had violent and treatment-resistant toxoplasma retinochoroiditis during pregnancy, leaving three of them with decreased visual acuity in spite of aggressive therapy. Termination of pregnancy appeared to help the recovery in two patients. Correspondence to C. Stephen Foster, MD, FACS 243 Charles Street Boston, MA 02114, USA Tel: Fax: fosters@uveitis.org Conclusion Pregnant state may provoke the recurrence of ocular toxoplasmosis. Key words: Toxoplasmosis, Retinochoroiditis, Uveitis, Pregnancy The authors have no proprietary or financial interest in any product described in this study. PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

30 TOXOPLASMOSIS is the most common cause of posterior uveitis in the world, accounting for over 80% of cases in some regions. 1 It is caused by the obligate intracellular protozoan Toxoplasma gondii. Recurrence of congenital toxoplasmosis is still the leading cause of Toxoplasma retinochoroiditis, but acquired ocular disease is more common than previously suspected. 2-5 More than 82% of congenitally infected individuals not treated as infants will develop retinal lesions by the time they reach adolescence. 6 Ocular toxoplasmosis is usually a recurrent disease, and two thirds of patients present with relapses. 7 We believe that recurrence is the result of release of Toxoplasma gondii trophozoites from cysts, which then actively infect the retina and release antigens stimulating an inflammatory retinochoroiditis. Some recurrences may actually be secondary to reinfection. In this article we report four patients who had recurrences of ocular toxoplasmosis during their pregnancies, suffered from visual loss and had treatment-resistant inflammation. METHODOLOGY We reviewed 1,243 charts of patients followed in the Uveitis Service at the Massachusetts Eye and Ear Infirmary from November 1986 to December One hundred ninety seven patients had a diagnosis of ocular toxoplasmosis, and 4 of these patients were pregnant at the time of recurrence of their ocular disease. Clinical records of these patients were reviewed for details of clinical presentation, ophthalmic history, complications and visual outcome at final follow-up. CASE REPORTS Case 1 A 22 year-old Hispanic woman, an immigrant from El Salvador, presented to the Ocular Immunology and Uveitis Service of the Massachusetts Eye and Ear Infirmary with complaints of a decreased visual acuity and redness in her right eye of five days duration. The patient had an episode of intense pain in the right eye two weeks prior to the visit. The pain persisted for one week. At the time of examination the patient was 28 weeks pregnant. She was taking prenatal vitamins and did not have any allergies. Her medical history was unremarkable. The patient s visual acuity was 20/50 (6/15) in the right eye and 20/20 (6/6) in the left eye. There were keratic precipitates on the corneal endothelium, and 3+ cells in the anterior chamber of the right eye. Dilated examination of the fundus revealed 3+ cells in the vitreous and an active focus of macula-threatening retinochoroiditis along the superotemporal arcade, and an old scar in the nasal periphery consistent with toxoplasmic lesions. After consulting with the patient s gynecologist, the patient was begun on clindamycin 300 mg orally three times a day and atovaquone 750 mg orally twice a day. On the follow-up visit, the patient s visual acuity in the affected eye had decreased to 20/300 (6/90). The amount of cells in the anterior chamber increased to 4+ and vitreous cells were 2+. Systemic prednisone 50 mg orally daily and topical prednisolone acetate, one drop every hour in the right eye, were started. In spite of this therapy, inflammation persisted for the next three months. The patient developed posterior vitreous detachment after two months of therapy during her 33rd week of pregnancy. Two weeks after giving birth, the patient returned for a follow-up visit. Examination revealed decreased vitreal cells to 1+ and mostly scarred toxoplasmic lesion. Visual acuity was 20/40 (6/12). Medications were discontinued. Six weeks after delivery the patient s visual acuity was 20/25 (6/7.5) in the affected eye and there was no active inflammation. The child was healthy. During her two-year follow-up the patient did not have recurrences of toxoplasmic retinochoroiditis. Case 2 A 24 year-old Hispanic female with a history of panuveitis in both eyes for four years, treated with periocular steroids with little success, was referred for evaluation. The patient s medical history was remarkable for 2 caesarean deliveries. The patient reported a history of recurrence of toxoplasmosis with her second pregnancy, but this did not affect the birth of a healthy child. At the time of the visit with us the patient s visual acuity was 20/30 (6/9) in the right eye and 20/70 (6/21) in the left eye. She was receiving one drop of prednisolone acetate hourly in the left eye. Slit-lamp examination showed posterior subcapsular cataract. Dilated fundoscopic examination revealed posterior vitreous detachment, 3+ old vitreous debris and 1+ cells in the left eye. There were bilateral choroidal lesions. Fluorescein angiography disclosed optic-disc staining of the left eye in the late frames. The patient had undergone an evaluation, which disclosed an antitoxoplasmic IgG antibody level of 2.2 EIA units (normal range 0-0.9). Chest X-ray, gallium scan, and other laboratory tests were normal. Because there were no vision-threatening lesions, we decided to observe the patient. Prednisolone acetate was slowly tapered. On subsequent visits the patient was found to have a progression of the cataract in her left eye and she scheduled phacoemulsification of the cataract combined with pars plana vitrectomy. Three weeks after the procedure, the patient developed a retinal detachment in the left eye and reactivation of uveitis. In that visit the patient also informed her care providers that she was 5 weeks PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

31 pregnant. The patient was referred for surgical treatment of the retinal detachment. The patient continued to have persistent inflammation in the left eye. Her condition finally stabilized with visual acuity of 20/50 (6/15) in the left eye by the third trimester of her pregnancy. The patient had a healthy baby and no reactivation of toxoplasmosis after the delivery. Case 3 A 20 year-old Hispanic female from Brazil with a history of ocular toxoplasmosis diagnosed at 13 years of age was referred for treatment of decreased visual acuity of her right eye and pain of one-week duration. The patient was 8 weeks pregnant at the time of her initial evaluation at the Ocular Immunology and Uveitis Service. The patient Table 1. Profile of 4 cases with recurrent ocular toxoplasmosis in pregnancy. Age Affected eye Visual acuity prior to flare-ups Weeks of pregnancy at the beginning of flare-ups Adverse events during flare-ups Total length of inflammation Worst visual acuity during flare-ups Final visual outcome History of flare-ups with previous pregnancies Case 1 22 OD 20/25 (6/7.5) 28 Posterior vitreous detachment 16 weeks 20/300 (6/90) 20/25 (6/7.5) No previous pregnancies Case 2 24 OS 20/40 (6/12) 2 Retinal detachment, cataract 24 weeks 20/200 (6/60) 20/50 (6/15) One flare-up Two previous pregnancies Case 3 20 OD 20/50 (6/15) 7 Cataract 35 weeks Light perception with correct light projection 20/50 (6/15) One flare-up Two previous pregnancies Case 4 26 OS 20/100 (6/30) 19 Progression of the disease to macula 9 weeks 20/600 (2/60) 20/80 (6/24) One flare-up One previous pregnancy Table 2. Medications used in the treatment of ocular toxoplasmosis. Medication Dihydrofolate reductase inhibitors Pyrimethamine Trimethoprim/sulfamethoxazole 1 Sulfonamides Sulfadiazine Sulfasoxazole Sulfadiazine/sulfamerazine/sulfamethazine ( triple sulfa ) Trimethoprim/sulfamethoxazole 1 Tetracyclines Minocycline Typical dose 100 mg loading dose mg daily for days 160mg/800 mg twice daily 1gm four times daily 160 mg/800 mg twice daily 100 mg twice daily Safety in pregnancy Category C* contraindicated in the first trimester; excreted in human milk Category C* contraindicated in third trimester and breastfeeding Category D* contraindicated in pregnancy and childhood Prednisone Macrolides Spiramycin Azithromycin Antiprotozoal Atovaquone Lincomycin derivative Clindamycin Folinic acid 1 gm/kg/day 500 mg three times daily for 3 weeks. May repeat after 21 days 500 mg daily for 3 weeks 750 mg four times daily for 4-6 weeks 300 mg four times daily for days 5-20 mg every day during pyrimethamine therapy Category C* Category B* Category B* Category B* Category A* * Grading system of medications in pregnancy: A = safety established in human studies, B = presumed safety based on animal studies, C = uncertain safety no human or animal studies; D = unsafe evidence of risk that may be justifiable in certain clinical circumstances; E = highly unsafe risk outweighs any possible benefit. 1 Trimethoprim/sulfamethoxazole is a combination of a dihydrofolate reductase inhibitor and a sulfonamide. 90 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

32 did not have any significant medical history or allergies, and was not taking any systemic medications. She was using topical prednisolone acetate hourly and scopolamine twice daily in her right eye. Her visual acuity was 20/400 (3/60) in the right eye and 20/20 (6/6) in the left eye. Slit-lamp examination revealed keratic precipitates on the corneal endothelium, 3+ cells in the anterior chamber, and pigment deposits on the anterior capsule of the lens of the right eye. Dilated fundus examination showed an active toxoplasmic lesion adjacent to an old scar in the right eye, and 3+ vitreous cells. The left eye had an old chorioretinal scar, but there were no signs of active inflammation. The patient was begun on atovaquone 750 mg orally twice daily and clindamycin 300 mg orally four times daily. Her follow-up visit in 10 days revealed worsening visual acuity in the affected eye: light perception with correct light projection, 3+cells in anterior chamber and increased vitreous cells (4+). Systemic prednisone was added to the regimen at a dose of 60 mg orally daily. In spite of this therapy, inflammation persisted for the remaining 7 months of the patient s pregnancy. On her follow-up visit 3 weeks post-delivery the patient s visual acuity remained at 20/400 (3/60). There was active inflammation with 3+ cells in the anterior chamber and extensive posterior synechiae obstructing the view of the fundus, along with posterior subcapsular cataract. Atovaquone was substituted with pyrimethamine 25 mg orally twice daily, folinic acid 5 mg orally twice weekly and sulfadiazine 1000 mg four times daily. The patient was continued on clindamycin. Control of inflammation was achieved within five weeks after delivery, with resulting visual acuity of 20/80 (6/24) in the right eye. Case 4 A 26-year-old Hispanic female with a history of ocular toxoplasmosis and glaucoma was referred for further evaluation and treatment. The patient was 22 weeks pregnant with her second child. She reported a history of recurrence of toxoplasmosis during her first pregnancy, leaving her with vision of 20/100 (6/30) in the left eye. The patient complained of loss of vision in her left eye for three weeks. She was started on therapy by the physician who saw her at that time. She was taking clindamycin 300 mg orally four times daily, sulfadiazine 500 mg orally twice daily, timolol maleate (Timoptic 0.5%, Merck Sharpe & Dohme, PA, USA) 1 drop and brimonidine (Alphagan Allergan, CA, USA) 1 drop twice a day in the left eye, as well as prednisolone acetate 1 drop four times daily in the left eye. Visual acuity was 20/25 (6/7.5) in the right eye and 20/600 in the left eye. Intraocular pressure was 15 mm Hg in the right and 21 mm Hg in the left eye. Slitlamp examination showed 2+ cells in the anterior chamber of the left eye. The dilated fundus examination was significant for an active inflammatory lesion in the macula and 2+ vitreal cells. Cup-to-disc ratio was 0.5 in the right and 0.6 in the left. An infectious disease specialist recommended increasing the dose of clindamycin to 450 mg orally four times daily, and increasing sulfadiazine to 1 g orally four times daily. Three weeks later the patient returned with resolution of the anterior chamber inflammation, but 2+ vitreous cells persisted. Control of inflammation was achieved in three more weeks of continued therapy. Visual acuity in the left eye remained 20/800 (1.5/60). RESULTS We have described four patients with recurrent ocular toxoplasmosis during their pregnancy. The clinical features are summarized in Table 1. Two patients had recurrence of toxoplasmic retinochoroiditis early in pregnancy, at 2 and 7 weeks, and the other two patients had recurrences at 28 and 19 weeks of pregnancy. All participants had a violent course of inflammation. Three patients had a prolonged course. Three of the four patients had permanent decrease in their best corrected visual acuity. Two developed cataracts and one developed retinal detachment. One patient had a decrease in visual acuity from 20/100 (6/30) to 20/800 (1.5/60) in the affected eye due to macular involvement in spite of aggressive therapy and a short duration of inflammation. Interestingly, two participants provided history of recurrences of ocular toxoplasmosis during their previous pregnancies. The details were unavailable for those incidents. Natural termination of pregnancy appeared to help the resolution of inflammation in two patients. DISCUSSION Many women are diagnosed with or experience recurrence of ocular toxoplasmosis during pregnancy. Newly diagnosed ocular toxoplasmosis in pregnant women is much less common than recurrence of toxoplasma retinochoroiditis. Pregnancy creates an interesting immunologic state in which strong immune responses are suppressed in order to prevent rejection of a fetus by the mother, and pregnancy-associated immunomodulation is thought to cause an ameliorating effect on some of the autoimmune diseases, such as multiple sclerosis, 8-11 rheumatoid arthritis, 12, 13 juvenile idiopathic arthritis 14 and sarcoidosis 15. The effect of pregnancy on the eye has not been studied well. Pregnancy associated immunomodulation may have a deleterious effect in patients with infectious ocular disorders. It is well known that immunocompromised patients are at increased risk for developing acute toxoplasmosis, which has a poor prognosis and may be rapidly fatal if left untreated. Ocular toxoplasmosis may follow a severe PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

33 course in patients with AIDS, having atypical features, such as large areas of retinal necrosis, 16 lesions arising perivascularly and not from old scars, 17 bilateral inflammation, and inflammation extending into the orbit and causing cellulitis and panophthalmitis. 18 Ocular toxoplasmosis may be more severe in elderly patients due to agerelated waning of host immune defenses. 19, 20 Certain timidity exists, even in specialty-trained uveitis specialists when treating pregnant patients with ocular toxoplasmosis. In one survey the majority of physicians stated that they would treat only severe vision-threatening acute toxoplasmosis in pregnant women. 21 Adverse effects which medications may cause to the fetus limit treatment options for pregnant women with toxoplasmic retinochoroiditis. In Table 2 we review the safety of medications used in treatment of ocular toxoplasmosis. Combination clindamycin and atovaquone may provide a safe approach to treatment of acute ocular toxoplasmosis in pregnant patients. The optimal duration of specific therapy must be adjusted according to therapeutic response. We define a positive response to treatment as a sharpening of the borders of retinochoroidal lesions and decrease of vitreal cells. We continue therapy in immunocompetent patients at least for days. Addition of oral corticosteroids may be started within 48 hours after initiation of antimicrobial therapy for intense ( 3+) vitreal cellular response. A macrolide, such as clindamycin or azithromycin, combined with atovaquone may provide an excellent treatment option for pregnant patients. Sulfonamides should be avoided in the third trimester, because they compete with bilirubin for serum proteins, causing kernicterus. Pyrimethamine is potentially teratogenic and should be avoided, especially in the first trimester. Comanagement with an infectious disease specialist is advised, as is consultation with an obstetrician. Some agree that there is a lack of evidence to support routine employment of antibiotic treatment for acute toxoplasmic retinochoroiditis in the general population. 22 But it may be especially important to treat medically pregnant women with ocular toxoplasmosis, since antibiotic therapy prescribed to mothers who have toxoplasmic retinochoroiditis during pregnancy decreases the percentage of children developing retinochoroidal toxoplasmic lesions during the 23, 24, 25, 26 first and second years of life. Pars plana vitrectomy may be useful in the care of selected patients with toxoplasma retinochoroiditis for removal of antigenic proteins, inflammatory cells, and persistent vitreous opacities. Intravitreal/intraocular antibiotic injections may be an option for treating pregnant patients with ocular toxoplasmosis, 27 thereby avoiding systemic toxicity, side effects, and possible teratogenicity. Martinez and collegues 27 reported a case series of pregnant women with active toxoplasmic retinochoroiditis, who were treated with a combination of intraocular clindamycin and dexamethasone with good visual outcomes. We have successfully treated several cases of ocular toxoplasmosis unresponsive to conventional therapy with intravitreal injections of clindamycin. 28 Patients in our series were not pregnant. They had intolerance to systemic therapy. The resolution of inflammation in these patients and improvement in visual acuity make us believe that penetration of systemic medications may be inadequate in treatment of ocular toxoplasmosis and is responsible for the poor visual outcomes in the pregnant women described in this article. The host immune system plays a critical role in response to ocular toxoplasmosis. It has been postulated that pregnancy may be a triggering factor in the recurrence of ocular toxoplasmosis Friedman and Knox 30 reported active toxoplasmic retinochoroiditis during five pregnancies of four patients, and Bosch-Driessen et al. 31 described seven women with ocular toxoplasmosis who had recurrences during pregnancy. Four of those women had reactivation of toxoplasmic lesions with each subsequent pregnancy. We believe that pregnancy not only predisposes a patient for recurrence of ocular toxoplasmosis, but may also present favorable conditions for a more aggressive form of this disease. Large cohort studies may lead to a better understanding of toxoplasmosis in pregnancy. References 1. Fernandes LC, Oréfice F. Aspectos clinicos e epidemiologicos das uveitis em service de referencia em Belo Horizonte Rev Bras Oftal 1996; 55: Silveira CM, Belfort R Jr, Burnier M, Nussenblatt RB. Acquired toxoplasmosis infection as the cause of toxoplasmic retinochoroiditis in families. Am J Ophthalmol 1988; 106: Glasner PD, Silveira C, Kruszon Moran D, et al. An unusually high prevalence of ocular toxoplasmosis in Southern Brazil. Am J Ophthalmol 1992; 114: Burnett AJ, Shortt SG, Isaac-Renton J, et al. Multiple cases of acquired toxoplasmosis retinitis presenting in an outbreak. Ophthalmology 1998; 105: Melamed JC. Acquired ocular toxoplasmosis-late onset. In: Nussenblatt RB, ed: Advances in Ocular Immunology: Proceedings of the 6th International Symposium on the Immunology and Immunopathology of the Eye. New York: Elsevier, 1994; Mets MB, Holfels E, Boyer KM, et al. Eye manifestations of congenital toxoplasmosis. Am J Ophthalmol 1996; 122: Da Mata AP, Oréfice F. Toxoplasmosis. In: Foster CS, Vitale AT, ed: Diagnosis and Treatment of Uveitis. Philadelphia: W.B. Saunders Company, 2002, pp Poser S, Poser W. Multiple sclerosis and gestation. Neurology 1983; 33: Frith JA, Mcleod JG. Pregnancy and multiple sclerosis. J Neurol Neurosurg Psychiatry 1988; 51: Birk K, Ford C, Smeltzer S, et al. The clinical course of multiple sclerosis during pregnancy and the puerperium. Arch Neurol 1990; 47: Confavreux C, Hutchinson M, Hours M, et al. Rate of pregnancy related relapses in multiple sclerosis. N Eng J Med 1998; 339: Da Silva JA, Spector TD. The role of pregnancy in the course and aetiology of rheumatoid arthritis. Clin Rheumatol 1992; 11: Ostensen M, Hugsby G. A prospective clinical study of the effect of pregnancy on rheumatoid arthritis and ankylosing spondylitis. Arthritis Rheum 1996; 26: Ostensen M. Pregnancy in patients with a history of juvenile rheumatoid arthritis. Arthritis Rheum 1991; 34: Mayock RL, Sullivan RD, Greening RR, Jones R Jr. Sarcoidosis and pregnancy. JAMA 1957; 164: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

34 16. Holland GN. Ocular toxoplasmosis in the immunocompromised host. Int Ophthalmol 1989; 13: Holland GN, Engstrom RE, Glasgow BJ, et al. Ocular toxoplasmosis in patients with the acquired immunodeficiency syndrome. Am J Ophthalmol 1988; 106: Moorthy RS, Smith RE, Rao NA. Progressive ocular toxoplasmosis in patients with acquired immunodeficiency syndrome. Am J Ophthalmol 1993; 115: Johnson MW, Greven GM, Jaffe GJ, et al. Atypical, severe toxoplasmic retinochoroiditis in elderly patients. Ophthalmology 1997; 104: Labalette P, Delhaes F, Margaron F, et al. Ocular toxoplasmosis after the fifth decade. Am J Ophthalmol 2002; 133: Holland GN, Lewis KG. An update on current practices in the management of ocular toxoplasmosis. Am J Ophthalmol 2002; 134: Stanford MR, See SE, Jones LV, Gilbert RE. Antibiotics for toxoplasmic retinochoroiditis: An evidence-based systematic review. Ophthalmology 2003; 110: Mets MB, Holfels E, Boyer KM, et al. Eye manifestations of congenital toxoplasmosis. Am J Ophthalmol 1996; 122: Couvrier J, Desmonts G, Thulliez P. Prophylaxis of congenital toxoplasmosis: Effects of spiramycin on placental infection. J Antimicrob Chemother 1998; 22(suppl B): Hohlfels P, Daffos F, Thulliez P, et al. Fetal toxoplasmosis: Outcome of pregnancy and infant follow-up after in utero treatment. J Pediatr 1989; 115: Daffos F, Forestier F, Capella-Pavlovsky M. Prenatal management of 746 pregnancies at risk for congenital toxoplasmosis. N Eng J Med 1988; 318: Martinez CE, Zhang D, Conway MD, Peyman GA. Successful management of ocular toxoplasmosis during pregnancy using combined intraocular clindamycin and dexamethasone with systemic sulfadiazine. Int Ophthalmol ; 22: Kump LI, Foster CS. Intraocular drug delivery in patients with uveitis under special circumstances. In press. Br J Ophthalmol. 29. Holland GN, O Connor GR, Belfort R, Remington JS. Toxoplasmosis. In: Pepose JS, Holland GN, Wilhelmus KR, eds. Ocular Infection and Immunity. St. Louis, Missouri: Mosby-Year Book, Inc, 1996; Friedman CT, Knox DL. Variations in recurrent active retinochoroiditis. Arch Ophthalmol 1969; 81: Bosch-Driessen LE, Berendschot TT, Ongkosuwito JV, Rothova A. Ocular toxoplasmosis: clinical features and prognosis of 154 patients. Ophthalmology 2002; 109: PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

35 VOL. 29 NO. 2 PHILIPPINE JOURNAL OF Ophthalmology APRIL CASE REPORT - JUNE 2004 Kristine T. Lo, MD Juancho Remulla, MD Alvina Pauline Santiago, MD Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Manila, Philippines Manifestations of Bardet-Biedl syndrome ABSTRACT Objective To report the first documented case of Bardet-Biedl syndrome at the University of the Philippines Philippine General Hospital. Methods This is a case report. Results A 7-year-old boy was diagnosed to have Bardet-Biedl syndrome based on the presence of five of the six primary manifestations of the disease: retinitis pigmentosa, obesity, postaxial polydactyly, learning disabilities, hypogenitalism, and renal dysfunction. Correspondence to Kristine T. Lo, MD Department of Ophthalmology and Visual Sciences University of the Philippines Philippine General Hospital Taft Avenue, Ermita 1000 Manila, Philippines Tel ext Fax kristinetlo@yahoo.com Conclusion Bardet-Biedl has ocular and systemic manifestations requiring a multidisciplinary approach to treatment. Keywords: Bardet-Biedl syndrome, Retinitis pigmentosa, Polydactyly, Hypogenitalism PHILIPP J OPHTHALMOL 2004; 29(2): PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004

36 BARDET-BIEDL syndrome (BBS) is an autosomal recessive condition characterized by rod-cone dystrophy, postaxial polydactyly, central obesity, mental retardation, hypogonadism, and renal dysfunction. Other features, though not always present, are hepatic fibrosis, diabetes mellitus, reproductive abnormalities, endocrine abnormalities, short stature, developmental delay, and speech deficits. 1 BBS is genetically heterogenous with at least 6 loci mapped to date. The most common gene is BBS1 located at chromosome Other genes identified are BBS2 at chromosome 16, BBS3 at chromosome 3, BBS4 at chromosome 15, BBS5 at chromosome 2, and BBS6 at chromosome The syndrome is relatively rare. It is estimated to be 1 in 160,000 in Switzerland. It is more common among Arab populations, specifically the Bedouins of Kuwait, 4 where interfamilial marriages are common. In Newfoundland, Canada, the incidence is 1 in 17,500. In the Philippines, no case has been reported yet. In this article we describe a case of Bardet-Biedl Syndrome presenting the features seen in the patient. CASE HISTORY A 7-year-old boy presented with a history of blurring of vision of two years duration, developmental delay, and nightblindness. Weighing 5.8 pounds at birth, he was born full term via spontaneous vaginal delivery to a 28- year-old primigravida, who was on rifampicin for pulmonary tuberculosis during the first 3 months of pregnancy. The patient was noted to have six digits on all extremities and a small penis. At 3 months of age, he was treated with rifampicin for primary complex. The eldest of 2 children (the sibling is normal), he was noted to have developmental delay upon consultation, with a developmental age equivalent to a 3- to 4- year-old. Best-corrected visual acuity was 20/70(6/21) on the right eye and 20/50 (6/15) on the left. Refraction was plano 4.00 cyl at axis 180 for both eyes. External adnexae were normal. Pupils were 3 mm and briskly reactive to light with no afferent pupillary defect (APD). The patient was orthophoric with a full range of ocular movements. Intraocular pressures were normal. Biomicroscopic examination revealed normal anterior segment. Indirect ophthalmoscopy revealed attenuated retinal vessels, atrophic retinal pigment epithelial (RPE) changes (Figure 1), and sparse pigmentation at the retinal periphery. Electroretinogram (ERG) showed findings consistent with tapetoretinal degeneration and retinitis pigmentosa (Figure 2). Visual-field examination was unreliable. Physical examination showed the patient was obese and had hexadactyly (Figure 3) and micropenis (Figure 4) measuring approximately 2 centimeters in length, with descended testes and a labia-like scrotum. The patient also had small dental roots and hypodontia (Figure 5). DISCUSSION Differential diagnoses Based on the ocular findings and systemic manifestations, the following differential diagnoses were considered: Weiss syndrome, Biemond II syndrome, Laurence Moon syndrome, Alstrom syndrome and Bardet- Biedl syndrome. Alstrom syndrome is characterized by tapetoretinal degeneration, obesity, preaxial polydactyly, diabetes mellitus, and neurogenic deafness. 5 The patient had only two (obesity, polydactyly) of these manifestations. His blood glucose was normal and he had no symptoms of diabetes. Neither did he have neurogenic deafness. The other findings seen in this patient (hypogenitalism, post axial polydactyly, and learning disability) are not manifestations of Alstrom syndrome. Biemond II syndrome is characterized by ocular defects (more specifically iris coloboma), learning disability, polydactyly, obesity, and hypogenitalism. 6 Although the patient had four of the main characteristics, retinal dystrophy is not a characteristic of this syndrome. Obesity, mental deficits, genital dystrophy, ner ve deafness, and short stature characterize Weiss syndrome, 6 which were all seen in the patient. But the patient had retinal dystrophy, which is not characteristic of this syndrome. Laurence Moon syndrome is characterized by retinitis pigmentosa or rod cone dystrophy, mental deficiency, hypogenitalism, and spastic paraparesis. 6 The patient, however, had no spastic paraparesis and patients with Laurence Moon syndrome do not show polydactyly and obesity. Similar to Laurence Moon syndrome, Bardet-Biedl syndrome is characterized by retinitis pigmentosa, obesity, postaxial polydactyly, learning disabilities, hypogenitalism, and in some cases, renal dysfunction. 6 With 5 of the 6 characteristics present in this patient, he was diagnosed to have Bardet-Biedl syndrome. Diagnostic criteria for Bardet-Biedl syndrome Beales et al. developed the diagnostic criteria for Bardet- Biedl syndrome: the presence of four primary features or a combination of three primary features plus two secondary features. 1 The primary features are rod-cone dystrophy, polydactyly, obesity, learning disabilities, hypogenitalism in males, and renal anomalies. The patient met 5 of these criteria. Secondary features are speech disorder/delay, strabismus, cataracts, astigmatism, syndactyly, brachydactyly, PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE

37 Figure 1. Fundus photos showing attenuation of retinal vessels and scattered RPE changes. Figure 4. Micropenis. Figure 2. ERG tracing. Figure 3. Hexadactyly. Figure 5. Hypodentia. developmental delay, polyuria and polydipsia. The patient had astigmatism, speech disorder, developmental delay, and brachydactyly (Figure 6). Additional secondary features are hypodontia/small tooth roots, dental crowding, high arched palate, mild spasticity, diabetes mellitus, left ventricular hypertrophy/ congenital heart disease, hepatic fibrosis, ataxia, poor coordination, and imbalance. The patient had hypodontia and small tooth roots. On the average, Bardet-Biedl syndrome is diagnosed at 9 years of age. The relatively late age of diagnosis can be explained by the slow development of clinical features. The first visual disturbance is usually night blindness. Retinopathy Retinitis pigmentosa is a disease characterized by progressive photoreceptor degeneration with RPE atrophy and intraretinal pigment migration of the RPE. 7 Symptoms include progressive contraction of the visual fields and night blindness. Decreased visual acuity can result from macular involvement of the disease. Figure 6. Brachydactyly. Attenuated vessels, mottling, granularity of the RPE, bone spicule formation, and waxy pale appearance of the disc are typical fundus findings in retinitis pigmentosa. Rod-cone dystrophy found in Bardet-Biedl syndrome is an atypical form of retinitis pigmentosa characterized by early decrease in visual acuity related to early macular involvement. The retinal periphery often has sparse bone spicule pigmentation, with central and peripheral RPE 96 PHILIPP J OPHTHALMOL VOL 29 NO. 2 APRIL - JUNE 2004